CA2831607C - Network transpose box and switch operation based on backplane ethernet - Google Patents

Network transpose box and switch operation based on backplane ethernet Download PDF

Info

Publication number
CA2831607C
CA2831607C CA2831607A CA2831607A CA2831607C CA 2831607 C CA2831607 C CA 2831607C CA 2831607 A CA2831607 A CA 2831607A CA 2831607 A CA2831607 A CA 2831607A CA 2831607 C CA2831607 C CA 2831607C
Authority
CA
Canada
Prior art keywords
network
tier
switches
connectors
transpose box
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CA2831607A
Other languages
French (fr)
Other versions
CA2831607A1 (en
Inventor
Alan M. Judge
Mark N. Kelly
Jagwinder Singh Brar
Michael David Marr
Daniel T. Cohn
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Amazon Technologies Inc
Original Assignee
Amazon Technologies Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Amazon Technologies Inc filed Critical Amazon Technologies Inc
Publication of CA2831607A1 publication Critical patent/CA2831607A1/en
Application granted granted Critical
Publication of CA2831607C publication Critical patent/CA2831607C/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/40Constructional details, e.g. power supply, mechanical construction or backplane
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/15Interconnection of switching modules
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/35Switches specially adapted for specific applications
    • H04L49/351Switches specially adapted for specific applications for local area network [LAN], e.g. Ethernet switches
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/70Virtual switches

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Small-Scale Networks (AREA)

Abstract

The deployment and scaling of a network of electronic devices can be improved by utilizing one or more network transpose boxes. Each transpose box can include a number of connectors and a meshing useful for implementing a specific network topology. When connecting devices of different tiers in the network, each device need only be connected to at least one of the connectors on the transpose box. One or more of the deployed electronic devices (e.g., switches, transpose boxes) in the network can transmit data based on a backplane Ethernet standard, such as 10GBASE-KR, 10GBASE-KX4, or 40GBASE-KR4.

Description

NETWORK TRANSPOSE BOX AND SWITCH OPERATION BASED ON BACKPLANE
ETHERNET
BACKGROUND
[0001] As an increasing number of applications and services are being made available over networks such as the Internet, an increasing number of content, application, and/or service providers are turning to networked and shared-resource technologies, such as cloud computing. Further, there is an increasing amount of data being stored remotely, such that data centers are increasingly expanding the amount of storage capacity and related resources.
A user or customer typically will rent, lease, or otherwise pay for access to.
resources through the cloud or across a network, and thus does not have to purchase and maintain the hardware and/or software to provide access to these resources.
[0002] In many instances, a customer will require more than one resource, such as a computing device, server, or other computing or processing device, to perform portions of an operation. As the number of customers increases, and the average number of resources per customer increases, there is a corresponding need to increase the available number of resources. In a data center context, this can mean adding many additional racks of servers. In order to accommodate the additional resources, the portion of the data center network that connects those resources to the external network needs to scale accordingly.
Such a network can require thousands of connections upon deployment, and the number can increase exponentially upon scaling to a larger deployment. In addition to the significant cost of purchase and installation, the large number of connections increases the likely number of connections that arc made incorrectly, and thus can affect the performance of the network.
SUMMARY
[0002a] According to one aspect of the present invention, there is provided a data transmission network, comprising: a first tier of network spine switches, each network spine switch in the first tier configured to receive and transmit data over a network; a second tier of network edge switches, each network edge switch in the second tier configured to receive and transmit data from at least one of the network spine switches in the first tier of network spine switches; and a network transpose box comprising a first group of network connectors associated with the first tier of network spine switches, a second group of network connectors associated with the second tier of network edge switches, and a meshing of communication media, the meshing of communication media being configured to implement a network topology that physically couples each one of the network connectors of the first group of network connectors individually to each one of the network connectors of the second group of network connectors to enable data transmission from the first tier of network spine switches to the second tier of network edge switches based on a backplane Ethernet standard, wherein a first network spine switch of the first tier of network spine switches is connected to a first network connector of the first group of network connectors of the transpose box via a single first cable, the first network spine switch being configured to transmit data with each one of the network edge switches in the second tier of network edge switches through the meshing of communication media.
10002b1 According to another aspect of the present invention, there is provided a data transmission system, the system comprising: a first tier of network spine switch devices; a second tier of network edge switch devices, the first tier of network spine switch devices and the second tier of network edge switch devices configured to transmit and receive data according to a backplane Ethernet standard; and a network transpose box configured to transmit data from the first tier of network spine switch devices to the second tier of network edge switch devices through a meshing of communication media, the network transpose box comprising: a first group of network connectors associated with the first tier of network spine switch devices, a second group of network connectors associated with the second tier of network edge switch devices, and a meshing of communication media, the meshing of communication media appearing as a single unified switch to one or more other network devices, wherein the meshing of communication media physically connects each of the first group of network connectors individually to each of the second group of network connectors and implements a network topology that connects, using a single transmission cable for each connection, multiple network spine switch devices of the first tier of network spine switch devices to multiple network edge switch devices of the second tier of network edge switch devices via the single transmission cable between each spine switch device of the first tier and l a each edge switch device of the second tier inside of the network transpose box, wherein a first network spine switch device of the first tier of network spine switch devices is connected to a first network connector of the first group of network connectors of the transpose box via a single first cable, the first network spine switch device being configured to transmit data with each one of the network edge switch devices in the second tier of network edge switch devices through the meshing of communication media.
10002c1 According to another aspect of the present invention, there is provided a method of deploying a data transmission network, comprising: selecting a network topology for at least a portion of the data transmission network; selecting one or more network devices to be deployed in the data transmission network, wherein the one or more network devices are configured to operate based at least in part on a backplane Ethernet standard;
based at least in part upon the selected network topology, selecting a network transpose box, the network transpose box including network connectors separated into a first group of network connectors associated with a first tier of network spine devices and a second group of network connectors associated with a second tier of network edge devices, and a plurality of communication media that is located within the network transpose box, the plurality of communication media having a meshing of communication media that implements the network topology that physically connects, using a single cable for each connection, multiple network spine devices of the first tier of network spine devices individually to multiple network edge devices of the second tier of network edge devices via the single cable between each spine device of the first tier and each edge device of the second tier inside of the network transpose box, the first group of network connectors each being connected to each of the second group of network connectors according to the selected topology using the meshing of communication media, the meshing of communication media appearing as a=single unified switch to one or more other network devices; and connecting a first network spine device of the first tier of network spine devices to a first network connector of the -first group of network connectors of the transpose box via a single first cable, the first network device being configured to communicate with each one of the network edge devices in the second tier of network edge devices through the meshing of communication media according to the selected network topology.
lb [0002d] According to another aspect of the present invention, there is provided a network transpose box comprising: a physical support structure; a set of network connectors supported by the physical support structure, the set of network connectors separated into a first logical group and a second logical group each containing a plurality of the network connectors, each network connector configured to accept a transmission cable for transmitting a signal; and a plurality of connection media that is located within the network transpose box, wherein the plurality of connection media facilitate signal transmission based on a backplane Ethernet standard, wherein each instance of the connection media connecting each of the network connectors of the first logical group to each of the network connectors of the second logical group enables a signal to be communicated between a transmission cable connected to any network connector of the first logical group and a transmission cable connected to any network connector of the second logical group, the connection media configured such that each network connector of the first logical group is physically connected individually to each network connector of the second logical group and each network connector of the second logical group is physically connected individually to each member of the first logical group, wherein at least some of a first tier of network spine switches is connected to at least one associated network connector of the first logical group of network connectors within the network transpose box using a single transmission cable and at least some of a second tier of network edge switches is connected to at least one associated network connector of the second logical group of network connectors using a single transmission cable within the network transpose box, and wherein the network transpose box is configured to transmit data from the first tier of network edge switches to the second tier of network edge switches through the plurality of connection media that is located within the network transpose box, and wherein the first logical group and second logical group appears to other network devices as a single unified logical group.
BRIEF DESCRIPTION OF THE DRAWINGS
100031 Various embodiments in accordance with the present disclosure will be described with reference to the drawings, in which:
1 c 100041 FIG. 1 illustrates an example environment in which various embodiments can be implemented;
[00051 FIG. 2 illustrates an example of a highly connected network design that can be used in accordance with various embodiments;
id [0006] FIG. 3 illustrates an example of a Clos network-style group of switches that can be utilized in accordance with various embodiments;
[0007] FIG. 4 illustrates an example of a group of switches utilizing a transpose box to make connections between tiers that can be used in accordance with at least one embodiment;
[0008] FIG. 5 illustrates the interior connections of an example transpose box that can be used in accordance with at least one embodiment;
[0009] FIG. 6 illustrates example keying approaches that can be used in accordance with various embodiments;
[0010] FIG. 7 illustrates an example process for utilizing a transpose box to make connections between tiers that can he used in accordance with at least one embodiment;
[0011] FIGS. 8(a)-8(d) illustrate approaches for sealing the number of network devices using a transpose box that can be used in accordance with various embodiments;
[0012] FIGS. 9(a)-9(e) illustrate approaches for deploying at least a portion of a network using one or more transpose boxes that can be used in accordance with various embodiments;
[0013] FIG. 10 illustrates an example process for scaling the number of network devices using a transpose box that can be used in accordance with various embodiments.
[0014] FIG. 11 illustrates an example configuration for connecting a network switch to a transpose box in accordance with various embodiments;
[0015] FIG. 12 illustrates an example of a three-tier Clos network-style group of switches including a logical-tier 2/3 switch that can be utilized in accordance with various embodiments;
[0016] FIG. 13 illustrates an example of a three-tier Clos network-style group of switches including a logical-tier 1/2 switch that can be utilized in accordance with various embodiments; and [0017] FIG. 14 illustrates an example of a three-tier Clos network-style group of switches including a logical-tier 1/2/3 switch that can be utilized in accordance with various embodiments.

DETAILED DESCRIPTION
100181 Systems and methods in accordance with various embodiments of the present disclosure may overcome one or more of the aforementioned and other deficiencies experienced in conventional approaches to deploying, connecting, maintaining, designing, and/or upgrading a network of electronic components. In computing networks such as data centers, for example, there are many levels (e.g., layers or tiers) of components and many connections between those levels. These can comprise, for example, a hierarchy of network switches connecting various host devices or other resources to an external network. The connections themselves can be made by any appropriate connection mechanism, such as fiber optic cable, network cable, copper wire, etc.
[0019] For each connection, a technician or other such person typically must connect the cable (or other connection mechanism) to one device, run the cable over a distance to another device, and connect that cable to the other appropriate device, Oftentimes these distances can be great, such that it can be easy to confuse the cables and end up making incorrect connections. Further, as networks such as data centers can have thousands of components, there is a likelihood that one or more cables will be installed incorrectly.
[0020] Further still, various network topologies require significantly more cabling than other topologies. For a high-radix network, for example, each device in a given tier may be fully connected to the devices in an adjacent tier, and there can be orders of magnitude more devices used in a high-radix network than for other types of networks. The result is that there are orders of magnitude more ports and connections, such that the number of cables needed can be significantly more than for other topologies, such as an oversubscribed hierarchical aggregation router pair network.
[0021] In various embodiments, a network transpose box or similar component can be used to facilitate the deployment, maintenance, and design of such a network. A
network box can include at least two logical sides, including one logical side for each tier or other set of components between which the transpose box sits. Each logical side can include an appropriate number of connectors, each able to accept a connection to a device of the appropriate tier.
[0022] The network transpose box also can include cabling, wiring, or other transmission media necessary to connect the connectors on each logical side of the transpose box. Instead of simple pass-through connections or one-to-many connections as in conventional
3 connection mechanisms, a transpose box can be designed in such a way that the transpose box itself implements a selected meshing or network topology. For example, in a Clos network where each switch of a first tier is connected to each switch of a second tier, the full meshing of connections can be handled inside the transpose box. In this way, for at least some transpose boxes, each switch only needs to run one connection (e.g., a multi-fiber cable) to the transpose box, instead of a number of connections necessary to connect to each device of the other tier. In other embodiments, a switch might have more than one connection to the transpose box (as may based at least in part upon factors such as cost, the selected network topology, the cabling technology, and the selected connection approach), but the number of overall cables is still significantly reduced from conventional cabling approaches. For example, the number of cables from a switch might be reduced from 24 or 48 cables to 4 cables or even a single cable, and those cables all go to a single location (e.g., the transpose box or set of transpose boxes) instead of to many different locations in a mesh or other topology. As should be apparent, reducing the number of connections that must be made by a technician to deploy such a network can significantly reduce the likelihood of a cabling error. Further, the reduction in cabling reduces the cost of the deployment, as well as the complexity and cost of scaling the network.
[0023] In some embodiments, the likelihood of a cabling error can further be reduced by keying, color coding, or otherwise uniquely identifying at least some of the connections to the transpose box. For example, each logical side of the transpose box can have connectors with a unique color or shape, to prevent a technician from connecting a cable to the wrong logical side (i.e., when all the connectors are on the same side of the transpose box). In transpose boxes that are fully meshed, it may not matter which connector the technician connects to, as long as the technician connects to the proper logical side. In other embodiments, various connectors might have specific keying when specific cables are to be connected to specific connectors. In some embodiments, the keying approach is tied to the network topology, and the number of unique types of keys can increase up to the number of possible types of connections for that topology, or the number of connectors on the transpose box. In some cases, each cable for a given type of connection may be uniquely keyed at each end such that the technician theoretically cannot improperly connect the devices (barring some problem with the cables themselves).
[0024] As discussed, the network topology can dictate the type of transpose box implemented in such a network. In some embodiments, the topology can be adjusted by replacing the transpose box. For example, a Clos network might have each device of two
4 tiers connected once to a Gins-meshing transpose box. If the network is to move to another topology, such as a dragonfly or butterfly topology, the technician can swap in an appropriate transpose box with the desired meshing, and reconnect each of the devices to the new transpose box. For complex topologies, the technician might connect multiple transpose boxes, each performing a portion of the meshing necessary for the selected topology.
100251 A network might also implement multiple transpose boxes for redundancy, such that if one transpose box fails the network can still function. Further, the redundancy allows one transpose box to be upgraded or otherwise modified or replaced without significantly affecting the availability of the network. For example, a network architect might want to increase the capacity of the network, and in some embodiments can replace an existing transpose box with a box having more connectors, in order to scale the network. Redundancy allows the box to be replaced without taking down the network.
[0026] In other embodiments, a network can use less than all available connectors on a transpose box upon initial deployment, such that at time of sealing the additional devices can connect to the available connectors. In other embodiments, additional transpose boxes can be added to the network, and connected to the existing transpose boxes in order to provide the desired meshing and/or connectivity.
[00271 In some embodiments, a network, including its various components can facilitate data transmission based on multiple transmission standards. For example, a network can include a portion that uses a backplane Ethernet standard, such as 10GBASE-KR, for the transmission of data.
100281 In particular, a switch in the network portion can be configured to transmit and receive data based on 1 OGBASE-KR operation. More specifically, the switch can include an edge connector through which data can be communicated using 10GBASE-KR. The edge connector can be connected to a transmission cable, which can in turn be connected to a transpose box. The transmission cable can, like the switch, be configured to facilitate communications based on I OGBASE-KR. Illustratively, the transmission cable can be implemented using copper or copper-based material that is capable of transmitting a 10613ASE-KR signal.
[0029] The transpose box to which the transmission cable is connected can additionally be configured to transmit and receive data based on 10GBASE-KR. In particular, the transpose box can include a suitable connector for interfacing with the transmission cable. The transpose box can additionally include components and/or circuitry for facilitating 10GBASE-KR communications. For example, the transpose box can include a set of copper traces capable of carrying a 10GBASE-KR signal.
[0030] Various other approaches can be used in accordance with the various examples and embodiments described below.
[0031] FIG. 1 illustrates an example of an environment 100 for implementing aspects in accordance with various embodiments. As will be appreciated, although a Web-based environment is used for purposes of explanation, different environments may he used, as appropriate, to implement various embodiments. The environment can include at least one electronic client device 102, which can include any appropriate device operable to send and receive requests, messages, or information over an appropriate network 104 and convey information back to a user of the device. Examples of such client devices include personal computers, cell phones, handheld messaging devices, laptop computers, set-top boxes, personal data assistants, electronic book readers, and the like. The network can include any appropriate network, including an intranet, the Internet, a cellular network, a local area network, or any other such network or combination thereof Components used for such a system can depend at least in part upon the type of network and/or environment selected.
Protocols and components for communicating via such a network are well known and will not be discussed herein in detail. Communication over the network can be enabled by wired or wireless connections, and combinations thereof. In this example, the network includes the Internet, as the environment includes a Web server 106 for receiving requests and serving content in response thereto, although for other networks an alternative device serving a similar purpose could be used as would be apparent to one of ordinary skill in the art.
[0032] The illustrative environment includes at least one application server 108 and a data store 110. It should be understood that there can be several application servers, layers, or other elements, processes, or components, which may be chained or otherwise configured, which can interact to perform tasks such as obtaining data from an appropriate data store. As used herein the term "data store" refers to any device or combination of devices capable of storing, accessing, and retrieving data, which may include any combination and number of data servers, databases, data storage devices, and data storage media, in any standard, distributed, or clustered environment. The application server can include any appropriate hardware and software for integrating with the data store as needed to execute aspects of one or more applications for the client device, handling a majority of the data access and business logic for an application. The application server provides access control services in cooperation with the data store, and is able to generate content such as text, graphics, audio, and/or video to be transferred to the user, which may be served to the user by the Web server in the form of HTML, XML, or another appropriate structured language in this example. The handling of all requests and responses, as well as the delivery of content between the client device 102 and the application server 108, can be handled by the Web server.
It should be understood that the Web and application servers are not required and are merely example components, as structured code discussed herein can be executed on any appropriate device or computing device as discussed elsewhere herein.
[0033] The data store 110 can include several separate data tables, databases, or other data storage mechanisms and media for storing data relating to a particular aspect.
For example, the data store illustrated includes mechanisms for storing production data 112 and user information 116, which can be used to serve content for the production side.
The data store also is shown to include a mechanism for storing log data 114, which can be used for purposes such as reporting and analysis. It should be understood that there can be many other aspects that may need to be stored in the data store, such as for page image information and access right information, which can be stored in any of the above listed mechanisms as appropriate or in additional mechanisms in the data store 110. The data store 110 is operable, through logic associated therewith, to receive instructions from the application server 108 or development server 120, and obtain, update, or otherwise process data in response thereto. In one example, a user might submit a search request for a certain type of item.
In this case, the data store might access the user information to verify the identity of the user, and can access the catalog detail information to obtain information about items of that type.
The information then can be returned to the user, such as in a results listing on a Web page that the user is able to view via a browser on the user device 102. Information for a particular item of interest can be viewed in a dedicated page or window of the browser.
[0034] Each server typically will include an operating system that provides executable program instructions for the general administration and operation of that server, and typically will include a computer-readable medium storing instructions that, when executed by a processor of the server, allow the server to perform its intended functions.
Suitable implementations for the operating system and general functionality of the servers are known or commercially available, and are readily implemented by persons having ordinary skill in the art, particularly in light of the disclosure herein.

[0035] The environment in one embodiment is a distributed computing environment utilizing several computer systems and components that are interconnected via communication links, using one or more computer networks or direct connections. However, it will be appreciated by those of ordinary skill in the art that such a system could operate equally well in a system having fewer or a greater number of components than are illustrated in FIG. 1. Thus, the depiction of the system 100 in FIG. 1 should be taken as being illustrative in nature, and not limiting to the scope of the disclosure.
[0036] An environment such as that illustrated in FIG. 1 can be useful thr an electronic marketplace or compute cloud, for example, wherein multiple hosts might be used to perform tasks such as serving content, executing large-scale computations, or performing any of a number of other such tasks. Some of these hosts may be configured to offer the same functionality, while other servers might be configured to perform at least some different functions. The hosts can be grouped together into clusters or other functional groups for the performance of specific tasks, such as may be provided as part of a data center, cloud computing offering, or processing service. The electronic environment in such cases might include additional components and/or other arrangements, such as those illustrated in the configuration 200 of FIG. 2, discussed in detail below.
100371 For example, FIG. 2 illustrates an example configuration 200 that represents a network design that can be used to route requests to specific host machines or other such devices, in order to provide users or applications with access to a variety of distributed resources. This example shows a typical design that can be used for a data center, wherein a source such as an end user device 202 or application 204 is able to send requests across a network 206, such as the Internet, to be received by one or more components of the data center. Properties of various components of the network, such as provisioned instances, etc., can be managed using at least one management system, component, or service 220. In this example, the requests are received over the network to one of a plurality of core switches 208, but it should be understood that there can be any of a number of other components between the network and the core switches as known in the art. As traditional differentiators have substantially disappeared, the terms "switch" and "router" can be used interchangeably.
For purposes of clarity and explanation this document standardizes on the term "switch," but it should be understood this term as used also encompasses routers and other devices or components used for such purposes. Further, the switches can include any appropriate switch, such as a multilayer switch that operates at different levels in an OSI (Open System Interconnection) reference model.

[0038] As illustrated, each core switch 208 is able to communicate with each of a plurality of aggregation switches 210, 212, which in at least some embodiments are utilized in pairs.
Utilizing aggregation switches in pairs provides a redundant capability in ease one or the switches experiences a failure or is otherwise unavailable, such that the other device can route traffic for the connected devices. As can be seen, each core switch in this example is connected to each aggregation switch, such that the tiers in this example are fully connected.
Each pair of aggregation switches 210, 212 is linked to a plurality of physical racks 214, each of which typically contains a top of rack (TOR) or "access" switch 216 and a plurality of physical host machines 218, such as data servers and other processing devices.
As shown, each aggregation switch can be connected to a number of different racks, each with a number of host machines. For the respective portion of the network, the aggregation pairs are also fully connected to the TOR switches.
[0039] As an additional benefit, the use of aggregation switch pairs enables the capability of a link to be exceeded during peak periods, for example, wherein both aggregation switches can concurrently handle and route traffic. Each pair of aggregation switches can service a dedicated number of racks, such as 120 racks, based on factors such as capacity, number of ports, etc. There can be any appropriate number of aggregation switches in a data center, such as six aggregation pairs. The traffic from the aggregation pairs can be aggregated by the core switches, which can pass the traffic "up and out" of the data center, such as back across the network 206. In some embodiments, the core switches are provided in pairs as well, for purposes including redundancy.
[0040] In some embodiments, such as high radix interconnection networks utilized for high-performance computing (UPC) or other such purposes, each physical rack can contain multiple switches. Instead of a single physical TOR switch connecting twenty-one hosts in a rack, for example, each of three switches in the rack can act as a local TOR
switch for a "logical" rack (a sub-rack of a physical rack or logical grouping of devices (hosts and/or switches) from multiple racks), with each local TOR switch connecting seven of the host machines. The logical racks can be implemented using physical or wireless switches in different embodiments. In some embodiments each of these switches within a high performance computing rack manages up to twelve servers, but the number can vary depending on factors such as the number of ports on each switch. For example, if a switch contains twenty-four ports, half of those ports typically will he host-facing and the other half will face the external network. A design in accordance with one embodiment could utilize seven racks with three switches in each, with each switch communicating (redundantly) with twelve servers, which would generally be equivalent to twenty-one separate racks each with a single TOR switch communicating with twelve servers, for example. In subsequent figures and description, it should be understood that physical or logical racks can be used within the scope of the various embodiments.
[0041] As discussed, the core switches in FIG. 2 are fully connected to the aggregation switches, and the aggregation switches are configured in pairs that are fully connected to a set of TOR switches. FIG. 3 illustrates an enhanced view of two such fully connected tiers of switches. The design presented illustrates a two-tier folded Clos network. As seen in the configuration 300 of FIG. 3, there are effectively two layers of switches: an upper tier or layer of spine switches and a lower tier or layer of edge switches. At least some of the edge switches (e.g., half of the edge switches in a traditional Clos), however, can be utilized as egress switches which pass data on to the network. The egress switches which logically sit at the "top" of the group of switches and pass data "up and out" of the group, such as to aggregation routers or other devices at a higher level tier, Each of the spine switches can be thought of as having a port out the logical "back" side to one of the egress switches, but the egress switches are simply selected from the forty-eight edge servers illustrated in a folded representation of FIG. 3. The egress switches simply have the only connections out of the group of switches, while the remaining edge switches have connections to underlying devices. All traffic into and out of the group of switches thus is routed through one of the three egress switches, although different numbers of switches can be used in different embodiments.
[0042] Even though the network may appear similar to the traditional core switch-based design of FIG. 2, the spine switches in this design function as core switches, but do not have any outbound connectivity. The layers of the group of switches have fully meshed connectivity, however, provided by the spine switches. The group of switches without the egress switches could function as a standalone network without any external connectivity.
Thus, some of the edge switches can be utilized as egress switches as illustrated. Otherwise, the fact that some of the edge switches are illustrated on the top layer and some on the bottom layer is meaningless from a network connectivity perspective with respect to the spine switches and the other edge switches, and there is very symmetric behavior.
The data within the group of switches can be pushed through a number of equidistant, fault-tolerant paths, providing the re-arrangably non-blocking behavior. With the paths being symmetric and equidistant, all the switches can follow the same routing protocol and spread the traffic evenly without a lot of overhead or additional logic. Further, the group of switches can be replicated multiple times within a data center, for example, wherein a Clos-style network effectively manages traffic across all of the groups in the data center.
[0043] Because the switches in the tiers of FIG. 3 are fully connected, such that each device on one tier is connected via at least one connection to each device in another tier, the number of cables needed to deploy such a design can be very large. For example, a single tier alone that contains 24 switches each with 48 ports would require 1,152 cables just to fully connect to the other tier. In a data center with many tiers and/or many more devices per tier, the number of cables quickly goes up to thousands or tens of thousands of cables. In addition to the expense of providing, installing, and maintaining these cables, there is a relatively high likelihood that at least some of the cables will be installed incorrectly. In the example above, 1,152 cables would require 2,304 individual connections. Even with a 99.9%
accuracy of installation, this would still result in a couple of connections being installed improperly.
Since many data centers run cables through walls, ceilings, floors, or other relatively hidden locations, the accuracy also can depend upon factors such as the labeling of the cables. Each additional step, however, introduces some additional likelihood for error in cabling. For example, if there is a 99.9% accuracy in labeling the cables and a 99.9%
accuracy in installing the cables, then there are now likely on the order of four cables that are installed incorrectly.
[0044] Further, networks such as those used in data centers often will need to scale over time to provide additional capacity. Using a design such as a high radix interconnection network design can require the number of switches to increase significantly each time the network is scaled, which not only can significantly increase the cost of the network but can also require an extensive amount of new cabling and re-cabling of existing devices. For example, horizontally scaling the deployment of FIG. 2 by adding another pair of core switches that have to be fully connected to twice the number of aggregation switches, which then must each be fully connected to a group of TOR switches as discussed in the example topology, can require a significant amount of work to connect the additional devices. This then further increases the likelihood of cabling errors, as each cable may need to be installed more than one time.
[0045] In some conventional networks, connection mechanisms exist that can simplify the cabling process. In one example, incoming fibers can be provided using a fiber bundle, which would require only a single connection for the bundle instead of a separate connection for each cable contained within that bundle. A connection mechanism can accept the fiber bundle on one side, and can connect each fiber within the bundle to a corresponding cable on the other side of the connection mechanism. These connection mechanisms are generally restricted to direct or straight pass-through connections, such that a first incoming fiber ("fiber #1") in the fiber bundle is connected to outgoing connector #1, incoming fiber #2 in the fiber bundle is connected to outgoing connector #2, and so on. Other mechanisms exist that accept a number of cables on one side (e.g., the incoming side) and connect each cable to a single corresponding connector on the other side (e.g., the outgoing side) of the connection mechanism. Such connection mechanisms have no real value in a fully connected network, however, where each switch in one tier is connected to each switch of another tier, which requires many more cables than are needed for single direct connections. There are no connection mechanisms used in conventional networks that provide the fully-connected design needed for high-radix designs and other network topologies as discussed herein, 10046] FIG. 4 illustrates an example configuration 400 that can be used in accordance with various embodiments, wherein connections between tiers in a network (e.g., between tiers) can be made using a transpose box 402 or similar network component. In this example, there are 24 spine switches in the upper tier and 48 switches in the lower tier, as in the example of FIG. 3. In the example of FIG. 3, however, each of the twenty-four upper tier switches 404 must be connected to each of the forty-eight lower tier switches 406, for a total of 1,152 cables or 2,304 individual connections that must be made, as discussed above.
In the example of FIG. 4, however, each of the twenty-four upper tier switches 404 need only be connected to the appropriate connector on the transpose box 402, resulting in twenty-four cables or forty-eight connections for the upper tier switches 404. The lower tier switches 406 also arc each connected only to the appropriate port on the transpose box 402, resulting in forty-eight cables or ninety-six connections for the lower tier switches 406.
Thus, by using the transpose box, the number of connections that need to be made to fully connect the tiers of switches is significantly reduced. In some cases, other types of cables (e.g., octopus cables, multi-ended cables, cables with multiple cores, etc.) or combinations of cables (e.g,, bundles of similar or different cables) can be used as well, while still obtaining a significant reduction in the amount of cabling and/or number of connections. In one specific example, an uplink cable to a transpose box might have 12 individual port connectors on one end (with two fibers each), and a 24 core trunk cable and a single 24-way connector at the transpose end. Many other variations are possible as well within the scope of the various embodiments.
[00471 The transpose box itself can be relatively small. In one example, a transpose box is about the size of a conventional switch, such as may have dimensions of about 19" wide and about 4"-5" deep, as may be able to fit within a conventional network rack.
Because fibers are small and flexible, and because the transpose boxes would in many cases be assembled on an assembly line or in a manufacturing facility, many fibers can be configured within a relatively small space. Further, since the transpose box is a self-contained component, there would be substantially no need for an outer protection layer on the fibers within a transpose box, such that even less room is needed for the full-connection design.
[0048] FIG. 5 illustrates a simplified example configuration 500 wherein six upper tier switches 502 arc fully connected to each of six lower tier switches 504 using a transpose box 506. While there arc an equal number of switches in each tier in this example for purposes of simplicity of explanation, it should be understood that there often will be unequal numbers of switches in different tiers, such as twice as many "lower tier" switches in folded Clos-based designs. The transpose box comprises some type of support structure, such as a frame, board, box, rack, enclosure, or other such structure or mechanism for supporting the a plurality of network connectors, each of which is able to receive a network cable for transmitting electronic, optical, or other such signals. As illustrated, the network connections can be arranged on different sides of the support structure, or can be at least partially on the same side or face but separated into different logical groups as discussed elsewhere herein. In some embodiments, the transpose box also can include circuitry and/or components for amplifying or transforming signals as discussed elsewhere herein.
[0049] Each of the upper and lower switches in this example can have at least six ports used to make connections to the transpose box 506, in order to make at least one connection for each of the switches in the other tier. It should be understood that the number of ports and/or switches can be different in other embodiments, as conventional switches can utilize as many as twenty-four or forty-eight ports for such connection purposes.
Because the transpose box 506 includes cables 516 or other connection mechanisms that provide connections to each of the switches in the other tier, there can be a single connector 510 for each of the upper tier switches 502 and a single connector 512 for each cable 514 connecting the transpose box 506 to one of the lower tier switches 504. In order to simplify cabling, the single cable between each switch and the transpose box can be a fiber bundle (as illustrated by the six individual fibers 518 shown to go into a cable 508) that includes at least one fiber for each connection to be made, such as at least one fiber for each switch in a given tier. In some embodiments, the fiber bundle will include a number of fibers equal to the number of ports on the switches in each tier (or at least a portion of the switches in a tier if different switches are used) such that if additional switches are added there is no need to replace the existing cables. in this example, if there are twenty-four ports on each switch and six switches in each tier, then a cable with twenty-four fibers would allow four individual connections to be made to each switch in the other tier (assuming a corresponding number of redundant connections within the transpose box itself).
F00501 As illustrated, each upper tier connector 510 is connected by at least one fiber (or other connection mechanism such as a wire or cable) to each lower tier connector 512, such that the connectors are fully connected. It should be understood that directional terms such as "upper" and "lower" are used for purposes of simplicity of explanation, and should not be interpreted as limiting the scope or implying any necessary orientation unless otherwise specified or suggested herein. Due to the fully-connected nature of the transpose box, each upper tier switch 502 will have a data transmission path to each lower tier switch 504, and vice versa, using only a single cable 508 between the upper tier switch 502 and the transpose box 506, along with a single cable 514 between the transpose box 506 and the target switch 504.
100511 In a simple approach, a single cable passes from each connector on the transpose box to a network component to be connected, such as a switch, server, or physical server rack. The interweaving of the transpose box provides for a meshing (e.g,, a full spread fan out or other topology) between any of the layers or tiers of the network, with only one cable (or two connections) per network device. In the event that a transpose box fails, the transpose box can simply be replaced with a different transpose box with at least the number of connections needed to be redone corresponding to, at most, the number of connectors on the transpose box, without any need for running new cable, rewiring, etc.
[00521 In some embodiments, there can be a different number of connectors on each logical "side" of the transpose box (e.g., "incoming" and "outgoing" sides, or a logical side facing a first tier and a logical side facing a second tier, logical north and south sides, etc.). It should be understood that these logical sides could actually correspond to any appropriate physical arrangement on the transpose box. An example transpose box could have n connections on one logical side and in connections on the other logical side, where each of the n incoming connections on one side is connected (singly or in blocks) to each of the in outgoing connections. In other examples, each logical outbound connection could be spread across multiple physical connectors, which could be less than the total number of available physical connectors. Various other topologies can be implemented as well. The transposing of the connections can be thought of similar to matrix multiplication, as there can be a matrix of outgoing connections represented as columns and incoming connections represented as rows.
In cases where fiber pairs are used for receive and transmit for each pair of connections (e.g., for optical transmission), each row and/or column could be further divided into pairs. The rollover or twisting of pairs of connections is handled within the transpose box, according to the selected matrix, as the rows are effectively converted into columns at the other side, and vice versa.
[0053] As discussed, such an approach is advantageous at least for the reason that reducing the number of cables reduces the cost of materials and the cost of deployment (i.e., making the physical connections). An example data center might have 80,000 cables between tiers, and the amount of necessary cabling is such that it typically is measured in tons of material.
As discussed above, reducing the amount of cabling can cut the cabling costs by as much as 90% or more, in addition to the savings obtained by using relatively small commodity switches instead of large network switches. On a per-port basis, such a deployment can run around twenty percent or less of the cost of a traditional large-scale network.
[0054] Another advantage is that a large reduction in the number of physical connections that must be made results in a corresponding reduction in the likely number of errors when making those connections. When deploying a conventional network, there is a significant operational cost and risk associated with the cabling, both in terms of properly installing the cabling and in maintaining the cabling (e.g., replacing cables when they fail). By utilizing one or more transpose boxes for interconnection, there is no need to connect a switch to every other switch in another tier, for example, but a single connection can be made to the appropriate transpose box from each switch (neglecting for the moment connections "up" to the network or connections to the host devices or other such components). The internal connections of the transpose box provide the full fan-out such that the connected switches will be fully connected between adjacent tiers. And because the transpose box performs the shuffling internally between the ports, cables such as multi-way optical cables can be used which include multiple optical fibers for providing transmit and receive data paths, instead of a large number of single pairs of fiber strands to provide the transmit and receive paths. For an optical cable with twenty-four internal fibers, for example, the twenty-four connections are virtually guaranteed to be correct (barring problems with the cable, for example) as long as the cable is attached to the correct connector on the transpose box.
[0055] In order to further reduce the probability of a cabling error for certain types of cabling, approaches in accordance with various embodiments can utilize one or more keying approaches to assist in connecting the cables to the appropriate connector.
For example a first keying approach 600 illustrates that each end of a cable can be a different color, such as by having a colored connector, a colored band near at least one end of the cable, etc. In one example, the end of each cable that is supposed to be connected to a switch could be a first color, and the end of the cable that is supposed to be connected to the transpose box could be a second color. Because the transpose box provides full connectivity, it does not matter in at least some embodiments which connector of a logical side of the transpose box the cable is connected to, and such a cabling approach could be used to ensure that each appropriate cable is connected to a switch at one end and a transpose box at the other.
[00561 In other embodiments, there might be cables with connectors of different colors to indicate whether the cable is going to a lower tier switch or an upper tier switch. For example, in FIG. 5 each of the lower tier switches 504 should be connected to one of the lower tier connectors 512 and not one of the upper tier connectors 510. In one example, each lower tier connector 512 is a color such as blue, and each cable from a lower tier switch 504 has a connector with a corresponding color, here blue, such that the person connecting the cable to the transpose box knows to connect the cable to the lower connector side of the transpose box. The upper tier connectors 510 can be a different color, such as red, so the person making the connections will be dissuaded from making an improper connection.
[0057] In some embodiments, the cables can have different keying approaches 620 instead of (or in addition to) different colors, such as a first keying approach 622 having a notch in a first location and a second keying approach 624 having a notch in a second location. By using different types of notches or other physical keys, cables cannot physically be connected to the wrong connector. Using the example above, each cable from a lower tier switch 504 might use the first keying approach 622, which ensures that the cable can only be connected to one of the lower tier connectors 512 if the upper tier connectors 510 use the second keying approach 624. It should be understood that in some embodiments all the connectors of the transpose box might be on the same side of the component, such that coloring or other distinguishing connector approaches can be further desirable.
[0058] In some cases, the deployment might require (or at least intend) that each switch be connected to a specific connector on the deployment box. in such an instance, there can potentially be a unique keying used for each switch within a selected group of switches. For example, a number of different types of keying 640 are shown in FIG. 6, including approaches with extension portions within the connector 642, approaches with notches or cut-outs 644, approaches with extension portions outside the connector 646, and/or approaches utilizing irregularly shaped connectors 648, such that each cable can only be connected to a particular switch and a particular connector on the deployment box. It should be understood that approaches can be reused for other groups of switches and/or other portions of the network where mis-cabling due to repetition of keying is at least highly improbable.
[0059] It also should be understood that while each connector shown in FIG. 6 might appear to have a single cable or fiber at the center, there can be many different configurations and types of connector. For example, a pair of fibers might result in side-by-side fiber endpoints, while a bundle of fibers might have several adjacent fibers within the same fiber bundle, or as portions of a single optical fiber. In other cases, each fiber might have a separate endpoint at the connector. MPO connectors, for example, can be used that are of different density, asymmetrical, or otherwise distinctive. There can be more fiber pairs, or more fiber cores, handled in one type of connector than another. The connectors can handle normal Tx/Rx fiber pairs, or any of a number of multi-path or multi-way fibers or cables. A
variety of other options also can be used as should be apparent.
100601 It also should be understood that while many examples provided herein relate to optical fibers and fiber-optic communications, approaches in accordance with various embodiments also can be used for other types of electronic signaling and/or data transfer as appropriate. For example, a transpose box can be used with electrical wiring, such as an active or passive transpose box for 10GBASE-T cable. In addition to providing the desired meshing, an active transpose box could also amplify or regenerate the signals in order to enable the signals to propagate over longer distances. Transpose boxes can also be used with twisted pair cables and a wide variety of communication or transportation media, such as I OGBASE-KR or 10GBASE-KX4, edge connectors, and custom cabling.
[0061] Further, other types of transpose boxes may be used that do not provide full meshing or full connectivity along a pure Clos-based design. For example, a transpose box might provide a specific number of straight pass-through connections (particularly for amplifying electrical signals). In other examples, the transpose box could be wired to help implement a different network topology, such as a dragonfly or butterfly network topology wherein a portion of the connections on one logical side of the box loop back to other connectors on the same logical side. In some examples, a cable out might have double capacity and accept information from two incoming cables. In some embodiments, the network topology can be selected and/or updated through selection of the transpose box to be implemented to perform the meshing.
[0062] Depending upon the type of cabling or other such factors, the type of connector(s) used also can help to ensure proper orientation of the cable at the connector.
For example, 10GBASE-T uses a single type of key for all connectors in order to ensure that the cable is installed with the proper orientation (such that each individual wire/fiber within the cable is connected to the appropriate location in the case of multiple wires/fibers being used). Such orientation-based keying can be used in combination with color-based keying, for example, to ensure that the cable is being connected at an appropriate location with the correct orientation. Various other keying approaches can be combined as well as discussed elsewhere herein. While two types of keying might be sufficient for a full fan-out from a north face to a south face of the box, the keying strategy can become more complicated as the complexity of the network topology increases. For example, a dragonfly network topology might utilize a local mesh and a global mesh, each with a distinct set of keying mechanisms.
Thus, the keying approach in at least some embodiments is selected based on the topology implemented, and there can be a number of key classes up to, and including, a unique key for each connector of the topology.
[0063] In some embodiments, a transpose box can enable different types of cables and/or connections on each logical side of the box. For example, the transpose box could contain circuitry and/or components to regenerate the signal received on one side for transmission using a different type of signal on the other side. In one example, optical fibers could be attached at a north side of the transpose box, with copper wire being attached at a south side of the transpose box, and the transpose box could perform the appropriate media conversion.
In a specific example, 1GBASE-T connections can be used between data servers and a transpose box, with fiber channels being output from the transpose box in order to provide for long distance communications (e.g., communications over the Internet or from a server room of a data center to the centralized network switches). Since fiber optics are currently much more expensive than copper wires, such an approach can provide the advantage that copper wire can be used to the extent possible, and then optical fibers used when necessary (with no meshing of those optical fibers being necessary, as the meshing is done by the transpose box, thus reducing the number of fibers needed). In some embodiments, a transpose box could even convert between physical and wireless connections, with each physical connection being meshed with an appropriate wireless signal or channel.

[0064] Another advantage to using transpose boxes as discussed herein is that the expertise and complexity of the network topologies is being centralized into the creation of the boxes.
By implementing such functionality, a data center technician does not need to understand the complexities of the various topologies, and instead only has to select and install the appropriate transpose box (which implements the appropriate topology).
Further, the transpose boxes can quickly and easily be tested during the manufacturing process (such as by ensuring that a proper signal is transmitted between appropriate connectors), such that there are no surprises or complex troubleshooting processes required for this portion of the network installation. If there is a network problem, a new box can be swapped in relatively quickly (e.g., on the order of eight minutes or less) to determine whether the box is the problem, as opposed to a lengthy process of testing all the individual cables and connections of the mesh. Such an approach also allows spare cabling to be run to the transpose box (instead of to the final destination) if all the ports are not to be used immediately. If one side (e.g., the north side) is fully wired to the existing infrastructure, scaling the network to add additional components at the south side then can be accomplished by connecting the new components directly to the transpose box.
[0065] FIG. 7 illustrates an example of a process 700 for deploying at least a portion of a network using at least one transpose box in accordance with at least one embodiment. In this example, a network designer, or other appropriate person first selects a type of network topology to be utilized for a specific network portion 702, such as a Clos-based portion as discussed above. Based on the selected topology, a corresponding transpose box is selected 704, as the meshing inside the transpose box implements that topology. The appropriate cabling for the transpose box (and number of network devices to be connected) is selected 706. As discussed, a transpose box can include different types of keying, and the number and types of cabling selected can depend upon factors such as the number and type of connectors on the transpose box. If the transpose box has different types of cabling on each logical side (e.g., optical fibers vs. copper wiring), then appropriate fibers, wires, or cables are selected and/or created and transceivers, media converters, or other necessary electronics can be inserted into the signaling path. Each first tier device (e.g., switch) is connected to an appropriate port on the first logical side of the transpose box 708, as may be dictated by the keying of the cabling and/or connector. Each second tier device (e.g., switch or network host) is connected to an appropriate port on the second logical side of the transpose box 710, as may be dictated by the keying of the cabling and/or connector. It should be understood that for this and other processes discussed herein, alternative, additional, or similar steps can be performed in similar or alternative orders, or in parallel, within the scope of the various embodiments unless otherwise stated. Once the devices of the selected tiers are connected to the transpose box, any remaining network components can be deployed 712 such that the network can be utilized for its intended purpose.
[0066] As mentioned, over time there often will be a need to scale or increase the size of the network deployment. In conventional systems, this often involves a significant re-cabling of the network. For example, if a group of switches has twenty-four first tier switches fully connected to twenty-four second tier switches, and twenty-four more switches are added to one of those tiers, then around 288 cables need to be moved and/or added just for that group alone. If, however, a transpose box capable of handling forty-eight switches on one logical side was used with a single cable connecting each device to the box, then only twenty-four new cables would need to be added as each additional device would simply need to be connected to the correct connection on the transpose box, Again, this is on the order of a 90% improvement over existing approaches. Further, a transpose box allows for incremental scaling that would be difficult using conventional approaches.
[0067] For example, FIG. 8(a) illustrates an example deployment 800 wherein there are three upper tier switches 802 fully connected via the transpose box 806 to each of three lower tier switches 804. In this example, a transpose box is initially (or subsequently) deployed that is able to handle more switches than are connected during the initial deployment. As can be seen, each of the devices is still fully connected to each device of the other tier.
[0068] If, for example, the network architect would like to scale to include an additional switch in each tier, the architect can direct the technician to add a switch to each tier, and connect each switch to the appropriate connector on the transpose box. As can be seen, the new upper tier switch 822 is connected to a corresponding connector on the upper side of the transpose box, and the new lower switch 824 is connected to a corresponding connector on the tower side of the transpose box. Because the transpose box in this example fully meshes the connectors on each side, each new switch is fully connected to all switches on the other side even though only a single additional cable was needed for each switch, In a conventional system (assuming adding a single switch to each tier is even an option), this would require at least one cable from each new switch to each switch of the other tier, for at least seven different cables (ignoring any issues with oversubscription, balancing, or other issues with other portions of the network).

[0069] As illustrated in the example configuration 840 of FIG. 8(e), the use of a properly meshed transpose box also can allow for asymmetric scaling of the network (where appropriate). In this example, the network can be scaled such that three additional lower tier switches 842 are added, such as to implement a three-stage, folded Clos network with a set of three spine switches connected between a set of lower tier switches (e.g., edge and egress) that is twice as large as the number of spine switches. In this example, each additional lower tier switch can be added using a single cable, while still being fully connected to the spine switches (upper tier).
[0070] In an even more asymmetric scaling example 860, FIG. 8(d) illustrates a configuration where switches 862 have been added to the lower tier, but the number of new switches is only a fraction of the number of upper tier switches 802. As can be seen, if the network allows it, switches can be added one or more at a time, to either side or both of the transpose box, Each additional switch only requires one cable to connect to the transpose box, with the connectivity being handled by the inner meshing of the transpose box 806.
[0071] Tn some instances, aspects such as cost or size limitations can prevent larger transpose boxes from being implemented initially, where at least a portion of the capacity of a transpose box will not be used right away. in some embodiments, the network architect can direct a technician to replace specific transpose boxes with larger boxes as needed. The network can then be scaled using any of the approaches discussed above. In other embodiments where there cost may prevent boxes from being swapped out unless there is a use for the old box, or for another such reason, it can be possible to introduce additional transpose boxes when scaling the network. Such an approach may not be optimal in all situations, as it can lead to network congestion and other such issues, but can be cheaper to implement and maintain in at least some situations.
[0072] For example, consider the example configuration 900 of FIG. 9(a). In this example, a single transpose box 906 is used to connect upper tier switches 902 and lower tier switches 904. Although in some embodiments all the connectors on the transport box are used, in this example there is one connector left available on each logical side of the transpose box. In the network is to be scaled such that two additional switches are to be added to each tier, then another transpose box 922 can be added as illustrated in the example 920 of FIG. 9(b). In this example, a second transpose box 922 is added to handle the additional switches. Because the transpose boxes are separate, however, there is no full connectivity between the switches connected to the first transpose box 906 and the switches connected to the second transpose box 922. In this example, an available connector on each of the upper and lower tiers of each transpose box can be used to provide at least one path 924 between the transpose boxes. The path can be implemented using at least one cable, fiber, bundle, additional transpose box, and/or any other appropriate communication or connection device. By linking the boxes in such a way, there is a path between each of the upper tier switches and each of the lower tier switches. As illustrated by the thicker lines in the figure, upper tier switch '2' can connect with lower tier switch '3' using the between-box connection path 924. Further, since the connectors used for the connection path were available, the boxes can be connected using a single cable or pair of cables in some embodiments. As discussed, the connection path 924 can be a point of congestion in certain systems, such that the approach might not be practical for certain implementations. An additional advantage of such an approach, however, is that there is no need to take down any functional part of the network while connecting the additional components. If a smaller network box is swapped out, for example, then the connectivity of that box will be unavailable for a period of the installation.
If the original transpose box and connectivity is untouched, however, then there is no such reduction in performance. The transpose boxes in such a situation thus also function as a safety zone when adding capacity. In some cases, the system could be designed with a certain size in mind, but only a portion implemented initially, such as every fourth transpose box at initial deployment. As the network scales, the additional transpose boxes (and other components) can be added as needed. Such an approach also can be used to connect two separate fabrics, as opposed to performing a traditional scaling operation. In such an approach, there would be no need to change any switches in either fabric, as long as the existing switches can be reconfigured as necessary.
[0073] In some embodiments, the additional transpose boxes might be deployed and used for redundancy and/or to prevent a single point of network failure for at least a portion of the network traffic. For example, FIG. 9(c) illustrates an example wherein there are three upper tier switches 902 and three lower tier switches 904 each fully connected by a first transpose box 942. The switches also are connected using a second transpose box 942.
Such a deployment provides for redundancy, and enables the network to remain functioning in the event of a problem or removal of one of the transpose boxes. In some embodiments, the redundancy can be built in as part of the design. In other embodiments, the second transpose box 942 can be used for redundancy when possible, and can be used for additional switches when scaling is desired, such as is described with respect to FIG. 9(b). In one embodiment, a group of switches with around twenty-four upper tier switches and around forty-eight lower tier switches might be connected by four transpose boxes (assuming a necessary number of ports, etc.) for purposes of redundancy, even though the group might not be able to further scale according to the current network design.
[0074] FIG. 10 illustrates an example process 1000 for scaling a network using transpose boxes that can be used in accordance with various embodiments. In this example, the initial stage of the network design is deployed, including at least one portion with at least one transpose box 1002. Once deployed, the network can operate as intended 1004.
At some point, a determination is made to scale up or increase some capacity of the network portion 1006. This determination can be made automatically in some embodiments, such as in response to the detection of capacity threshold being reached or being projected to be reached, and/or can be made manually, such as where a network administrator indicates that capacity is to be increased. Many other such determinations can be made and/or passed along as should be apparent.
100751 In at least some embodiments, a determination is first made as to whether there is an appropriate number of available connections, of the necessary type, to handle the increase 1008. If so, the additional devices (e.g., switches or hosts) or network portions can be connected to the available connectors on the transpose box 1010, and the expanded network can operate as intended, if there are not a sufficient number of available connectors, a determination can be made in at least some embodiments whether an upgraded box is available and/or allowed to be installed in the network portion 1012. If an upgraded box is available and allowable, the transpose box can be replaced with the larger box (at least in terms of connections and not necessarily size) 1014, and the additional devices can be connected as desired. If a larger box cannot be utilized, at least one additional transpose box can be added to the network portion 1016 and the transpose boxes can be connected 1018 as necessary per the selected network topology. As discussed above, the new or additional boxes can be selected at least in part based upon the network topology in addition to the number and/or type of necessary connections. It should be understood that a similar process can be used when the network topology is changed, where additional or alternative transpose boxes are selected to implement the new topology. Further, in some embodiments there can be multiple levels of transpose boxes between network components in order to implement complex topologies.
100761 In some embodiments, one or more switches and/or transpose boxes can be implemented using an extension of a network standard, such as the IEEE 802.3ap standard, for example, which provides Ethernet operation for backplane applications (e.g., blade servers, routers/switches with upgradeable line cards, etc.). The IEEE 802.3ap standard, in general, provides for two different 10Gbit/s implementation standards: 10GBASE-KR and 100BASE-KX4. Implementations based on I OGBASE-KR operate over a single lane and use the same physical layer coding as I OGBA SE-LR/ER/SR. By contrast, implementations based on 10GBASE-KX4 operate over four lanes and use the same physical layer coding as I OGBASE-CX4. Data transmission utilizing either implementation standard, however, is conventionally performed using copper traces embedded on printed circuit boards over very short distances (e.g., traces running within a single deployment chassis).
[0077] According to some embodiments, one or more switches and/or transpose boxes can be configured to operate based on 10GBASE-KR and/or I OGBASE-KX4 standards.
The one or more switches and transpose boxes can further be connected to one another without using a baekplane type configuration. More specifically, connections between the switches and transpose boxes need not be limited to copper traces embedded on printed circuit boards.
Rather, connections between the devices can be facilitated using a variety of different connection mechanisms, such as through the use of a set of transmission cables.
[0078] Through implementing switches and transpose boxes that leverage 10GBASE-KR
and/or I OGBASE-KX4, embodiments enable low cost operation with reduced complexity. In particular, many switches currently manufactured and sold by various networking vendors include unnecessary hardware components and configurations that add to the cost and complexity of the switches. By enabling switches and transpose boxes to communicate using I OGBASE-KR and/or 10GBASE-KX4, embodiments enable certain hardware components to be removed from the switches. As a result, switches can be constructed that are low in cost, consume less energy, have improved thermal characteristics, and provide for greater space efficiency.
[00791 FIG. 11 illustrates an example configuration 1100 including a switch connected to a transpose box 1116 via a transmission cable 1114 in accordance with one embodiment. Switch 1102 can include a printed circuit board 1104, which, in turn, includes processor 1106, one or more application specific integrated circuits (ASICs) 1108, and an edge connector 1110. Switch 1102 can also include other components and circuitry (not shown) for enabling network switching and other data processing. Such components and circuitry are well known and need not be specifically discussed.

[0080] In some embodiments, switch 1102 can be configured to operate based, at least in part, on 10GBASE-KR. As a result, switch 1102, in certain embodiments, need not include additional components and/or circuitry for separately generating a signal capable of being transmitted through transmission cable 1114 to transpose box 1116. In some embodiments, switch 1102 can be configured to additionally operate based, at least in part, on a non-backplane Ethernet standard, In such configurations, switch 1102 can include a first set of ports that are connected to transpose box 1116 and facilitate data transmission using 10GBASE-KR. Switch 1102 can additionally have a second set of ports that are connected to external devices (e.g., any device capable of connecting to a network) and facilitate data transmission using a non-backplane Ethernet standard and/or copper-based media (e.g., 10GBASE-T, fiber optics, etc.). In some embodiments, switch 1102 can include any suitable circuitry and/or components (such as appropriate PHY chips, transceivers, edge ports, etc.) to convert signals.
[0081] In some embodiments, switch 1102 can be fully or partially enclosed within a housing or chassis. In some cases, the chassis can include an opening, access panel, door, or other suitable mechanism for facilitating the connection of switch 1102 to transpose box 1116 using transmission cable 1114. For example, the chassis might provide an opening that exposes edge connector 1110, and enables a transmission cable to be run from edge connector 1110 to transpose box 1116.
[0082] Processor 1106 can be implemented as one or more integrated circuits.
For example, processor 1106 can be a microprocessor, a mierocontroller, and/or the like.
Processor 1106 can be configured to receive inbound data, perform switching functions (e.g., determine a physical path for data), transmit outbound data, etc. In some embodiments, processor 1106 can be configured to execute machine-readable instructions stored on a storage unit (not shown) accessible by the processor. The storage unit can include any suitable volatile and/or non-volatile storage mediums including suitable variations of random access memory (RAM), read-only memory (ROM), hybrid types of memory, storage devices, hard drives, optical disc drives, etc.
10083] The one or more ASICs 1108 can be implemented as one or more integrated circuits. In some embodiments, ASICs 1108 can be programmed and/or configured to perform certain types of processing. For example, ASICs 1108 can be configured, in some embodiments, to enable the transmission of data based on the 10GBASE-KR
standard.
Illustratively, ASICs 1108 can be configured to initiate the generation of data signals in accordance with 10GBASE-KR. While the example shown in FIG. 11 uses one or more ASICs, any suitable network switching chips can be used.
[0084] Edge connector 1110 can be situated along an edge of printed circuit board 1104.
Edge connector 1110 can include a set of printed circuit board traces that terminate at or near the edge of printed circuit board 1104. The set of traces can be implemented and configured in a manner as to enable data transmission based on 10GBASE-KR. For example, the set of traces can be implemented, at least in part, using suitable copper or copper-based material. In some embodiments, edge connector 1110 can be suitably connected to the one or more ASICs 1108. Illustratively, one or more of the set of printed circuit board traces included in edge connector 1110 can be directly or indirectly connected to one or more pins of the one or more ASICs 1108. In some embodiments, edge connector 1110 can be a 12-way or 24-way connector. That is, edge connector 1110 can include 12 or 24 different data transmission paths (e.g., 12 or 24 ports).
[0085] In some embodiments, edge connector 1110 can be configured to connect with transmission cable 1114. In particular, edge connector 1110 can be adapted to receive and electrically couple with edge connector socket 1112 of transmission cable 1114. For example, edge connector 1110 can be configured and shaped to securely' interface with edge connector socket 1112. Edge connector socket 1112 can be, for example, a female type electrical connector.
[0086] While the embodiment described herein includes an edge connector situated on a printed circuit board connected to an edge connector socket of a transmission cable, any suitable connection mechanism known in the art can be used for interfacing switch 1102 with transmission cable 1114. For example, switch 1102 can include an edge connector socket directly embedded on its printed circuit board. The transmission cable can, in turn, include a suitable connector for interfacing with the edge connector socket.
[0087] Transmission cable 1114 can be implemented as a suitable type of cabling or wiring for connecting switch 1102 to transpose box 1116 (e.g., via interfacing with connectors in the switch and box), and facilitating data transmission based on 10GBASE-KR. For example, transmission cable 1114 can be a ribbon cable, a cable based on twisted pairs, a passive cable, and/or a similar cable with sufficient channel bandwidth for 10GBASE-KR
operation. The cable can further include suitable copper or copper-based wiring (e.g., wiring that maintains signal fidelity within 10GBASE-KR specifications). In some embodiments, the transmission cable can be flexible, as is the case with various types of ribbon cables known in the art. As shown in FIG. 11, transmission cable 1114 can include edge connector socket 1112 situated at one end. On the opposite end, transmission cable 1114 can include a suitable connector for interfacing with transpose box 1116. In some embodiments, transmission cable 1114 can include a number of data transmission paths at least equal to the number of ports that is to be connected between switch 1102 and transpose box 1116. In order to provide protection from external electromagnetic interference, transmission cable 1114 can be shielded using a suitable material. It should be appreciated that while only one transmission cable 1114 is shown as connecting switch 1102 and transpose box 1116, any number of transmission cables can be used. In some embodiments, the number of transmission cables used can be significantly less than the number of ports of switch 1102 that are to be connected to transpose box 1116.
[00881 Transpose box 1116 can include any suitable circuitry and/or components for supporting data transmission based on 10GBASE-KR. For example, transpose box 1116 can include a printed circuit board with a set of copper traces and an appropriate set of connectors for connecting to one or more switches via transmission cables. Each connector can be a 12-way or 24-way connector. In some embodiments, transpose box 1116 can further include at least two logical sides, with each side configured to connect to a different tier of switches. In some embodiments, the different logical sides of transpose box 1116 can be connected via the aforementioned set of traces. In doing so, a switch connected to the transpose box in one tier can transmit data to a switch connected to the same box in another tier, In some embodiments, transpose box 1116 can further include circuitry and/or components for amplifying signals. For example, transpose box 1116 can include circuitry that amplifies signals so that transmission of data can extend further than the current distance limitations of the I OGBASE-KR standard. For example, circuitry included in transpose box 1116 might amplify a signal so that data can be transmitted up to 10 or more meters, 100891 Set of connection 1118 can be implemented using any suitable connection mechanism for connecting switch 1102 to external devices (e.g., external switches, servers, edge hosts, outside connections, etc.). As discussed, data transmission over the connections to the external devices can be, in some embodiments, facilitated using a non-backplane Ethernet standard.
[0090] FIG. 12 illustrates an example configuration 1200 of a three-tier Clos-based network that includes a logical-tier 2/3 switch. In some embodiments, a software program executed by a processor of a network management server (not shown in FIG. 12) can operate the logical-tier 2/3 switch in such a manner so as to make the logical switch appear to other network devices as if it were a single unified switch. As seen in the configuration, there are effectively three layers of switches: an upper tier 1202, middle tier 1206, and lower tier 1208.
In some embodiments, data transmission over at least a portion of the network can be based on 10GBASE-KR. For example, the various network components can include ASICs, copper traces, wires, and/or cables that are suitable for facilitating communication using the 10GBASE-KR standard. Those paths of the network communicating using 10GBASE-KR

are shown as dashed lines in FIG. 12.
[0091] In the example of FIG. 12, the connections between the upper and middle tiers are made using transpose box 1204. More specifically, each switch in upper tier 1202 is shown as being connected to a first logical side of transpose box 1204. Each switch in middle tier 1206 is additionally shown as being connected to a second logical side of transpose box 1204.
Through connecting the upper and middle tier switches in this manner, data transmission paths are established between each switch in the upper tier and each switch in the middle tier.
In some embodiments, the switches and transpose box can be located close to each other, such as in the same or in nearby network equipment racks.
[0092] In one embodiment, each switch in upper tier 1202 can have components and a configuration similar to switch 1102 illustrated in FIG. 11. For example, each switch can include a processor, one or more ASICs, and an edge connector. These components can enable a switch to be configured to operate based on 100BASE-KR. In some embodiments, the edge connector of each switch can be connected to a transmission cable, which in turn connects to transpose box 1204. The transmission cable between each switch in the upper tier and transpose box 1204 can be up to 10 meters in length. The transmission cable can further include a number data transmission paths equal to the number of ports of an upper tier switch that are to be connected to transpose box 1204. As discussed, by bundling the different transmission paths into a single cable, and by leveraging a transpose box, deployment of a network can be made simpler and more time efficient.
[0093] Transpose box 1204 can have components and a configuration similar to transpose box 1116 illustrated in FIG. 11. In particular, transpose box 1204 can include a printed circuit board embedded with a set of copper traces. The traces can be configured in a manner (e.g., with sufficient bandwidth characteristics, etc.) such that data transmission can be facilitated based on 10GBASE-KR. On one logical side, transpose box 1204 can connect to one or more transmission cables connected to the switches in upper tier 1202.
On a second logical side, transpose box 1204 can connect to one or more transmission cables connected to the switches in middle tier 1206. By connecting with the switches of the upper and middle tier in this manner, transpose box 1204 can enable each switch in the upper tier to have a connection with each switch in the middle tier.
[0094] In some embodiments, transpose box 1204 can additionally include circuitry and/or components for amplifying signals. For example, transpose box 1204 can include circuitry that amplifies signals so that the transmission of data can be extended to greater distances.
100951 As noted previously, one or more transmission cables can connect transpose box 1204 to the switches in middle tier 1206. The transmission cables can be, in some embodiments, up to 10 meters in length. The transmission cables can additionally include a number data transmission paths equal to the number of ports of a middle tier switch that are to be connected to transpose box 1204. In some embodiments, the transmission cables can be substantially similar to the transmission cables connecting transpose box 1204 with the switches in upper tier 1202. In certain embodiments, the transmission cables can include features or characteristics that distinguish the cables from the cables connecting transpose box 1204 to the switches in upper tier 1202. For example, the cables might differ in color, connector shape, connector type, size, width, and/or the like. By using transmission cables with different characteristics, a network technician can more easily determine which cables connect to switches in the middle and upper tiers.
[0096] In some embodiments, each switch in the middle tier can have components and a configuration similar to switch 1102 illustrated in FIG. 11. For example, each switch in the middle tier can include a processor, one or more ASICs, and an edge connector.
These components can enable a switch to be configured to operate based on 10GBASE-KR. In some embodiments, the edge connector of each switch can be situated on a first logical side attic switch. The edge connector on the first logical side can be connected to a transmission cable (as discussed above), which can in turn be connected to transpose box 1204. In some embodiments, each switch in the middle tier can additionally include circuitry and/or components that enable data transmission with one or more switches in the lower tier 1208 of FIG. 12. For example, the switches in the lower tier might be configured to perform data transmission based on other transmission media and/or standards, such as fiber-optics, 10GBASE-T, etc. As such, a switch in middle tier 1206 can include circuitry and/or components (such as appropriate PHY chips, transceivers, edge ports, etc), to convert signals and enable data transmission with the switches in the lower tier. Due to such a configuration, each switch in middle tier 1206 can have a portion (e.g., half) of its ports configured based on the 10GBASE-KR standard and another portion (e.g., the other half) configured based on a different data transmission media and/or standard (e.g., fiber-optics, 10GBASE-T, etc.).
[0097] As discussed, the network configuration illustrated in FIG. 12 facilitates data transmission between upper tier 1202, transpose box 1204, and middle tier 1206 based on 10GBASE-KR, Through enabling data transmission between the second and third tier switches based on 10GBASE-KR, the overall cost and energy consumption in constructing and maintaining the Clos network can be reduced. In particular, the 10GBASE-KR
standard, relative to other standards, is less costly to implement since fewer components are used.
Thus, by extending (e.g., by using transpose boxes, edge connectors and transmission cables) the 10GBASE-KR standard for use outside of a conventional backplane scenario, embodiments enable the deployment of networks that are less costly and more energy efficient.
100981 FIG. 13 illustrates an example configuration 1300 of a three-tier Clos network including a logical-tier 1/2 switch. In some embodiments, a software program executed by a processor of a network management server (not shown in FIG. 13) can operate the logical-tier 1/2 switch in such a manner so as to make the logical switch appear to other network devices as if it were a single unified switch. As seen in the configuration, there are effectively three layers of switches: an upper tier 1302, a middle tier 1304, and a lower tier 1306. In some embodiments, data transmission over at least a portion of the network can be based on 10GBASE-KR operation. For example, various network components can include ASICs, copper traces, wires, and/or cables that are suitable for facilitating communication using the 10GBASE-KR standard. Those paths of the network communicating using 10GBASE-KR are shown as dashed lines in FIG. 13.
[0099] In the example of FIG. 13, switches 1308 and 1310 located in the middle tier are connected to switches 1314 and 1316 located in the lower tier via transpose box 1312. While only two switches in the middle tier and two switches in the lower tier are shown as being connected using a single transpose box in FIG. 13, any number of switches can be connected using any number of transpose boxes. In some embodiments, the switches and transpose box can be located close to each other, such as in the same or in nearby network equipment racks.
[0100] As shown in configuration 1300, switches 1308 and 1310 of middle tier 1304 can be connected, on one logical side, to one or more switches in upper tier 1302.
Data transmission between the middle tier switches and the upper tier switches can occur using any suitable transmission media and/or standard. For example, data transmission between middle tier switches 1308 and 1310 and the upper tier switches can be based on fiber-optics or I OGBASE-T. As such, middle tier switches 1308 and 1310 can include circuitry and/or components (such as appropriate PHY chips, transceivers, edge ports, etc), to convert signals and enable data transmission with the switches in the upper tier.
[0101] Switches 1308 and 1310 can additionally be connected to transpose box 1312 on a second logical side. In particular, middle tier switches 1308 and 1310 can include components and configurations similar to that of switch 1102 illustrated in FIG. 11. For example, switches 1308 and 1310 can each include a processor, one or more ASICs, and an edge connector. These components can enable a switch to be configured to operate based on 10GBASE-K R. In some embodiments, the edge connector of each switch can be connected to a transmission cable, which in turn connects to transpose box 1312. In some embodiments, the transmission cable between each of the switches 1308 and 1310 and transpose box 1312 can be up to 10 meters in length. The transmission cable can further include a number data transmission paths equal to the number of ports of switch 1308 or 1310 that are to be connected to transpose box 1312. Due to such configurations, switches 1308 and 1310 can each have a portion (e.g., half) of its ports configured based on 10GBASE-KR
and another portion (e.g., the other half) configured based on a different data transmission media and/or standard (e.g., fiber-optics, 10GBASE-T, etc.).
[0102] Transpose box 1312 can have components and a configuration similar to transpose box 1116 illustrated in FIG. 11. In particular, transpose box 1312 can include a printed circuit board embedded with a set of copper traces. The traces can be configured in a manner such that data transmission can be facilitated based on the 100BASE-KR
standard. On one logical side, transpose box 1312 can connect to one or more transmission cables connected to switches 1308 and 1310. On a second logical side, transpose box 1312 can connect to one or more transmission cables connected to switches 1314 and 1316 of lower tier 1306. By connecting with the switches in this manner, transpose box 1314 can enable switches 1308 and 1310 to have a connection with switches 1314 and 1316.
[0103] In some embodiments, transpose box 1312 can additionally include circuitry and/or components for amplifying signals. For example, transpose box 1312 can include circuitry that amplifies signals so that transmission of data can be extended to greater distances.
[0104] As noted previously, one or more transmission cables can connect transpose box 1312 to lower tier switches 1314 and 1316. The transmission cables can be, in some embodiments, up to a length of 10 meters. The transmission cables can additionally include a number data transmission paths equal to the number of ports of a lower tier switch that are to be connected to transpose box 1312. In some embodiments, the transmission cables can be substantially similar to the transmission cables connecting transpose box 1312 with middle tier switches 1308 and 1310. In certain embodiments, the transmission cables can include features or characteristics that distinguish the cables from the cables connecting transpose box 1312 to middle tier switches 1308 and 1310. For example, the cables may differ in color, connector shape, connector type, size, width, and/or the like.
101051 Switches 1314 and 1316 can each have components and a configuration similar to switch 1102 illustrated in FIG. 11. For example, each switch can include a processor, one or more ASICs, and an edge connector. These components can enable a switch to be configured to operate based on 10GBASE-KR. In some embodiments, the edge connector of each switch can be connected to a transmission cable, which in turn connects to transpose box 1312. In some embodiments, switches 1314 and 1316 can each be connected to external devices (e.g., external switches, servers, edge hosts, outside connections, etc.). For example, FIG. 13 shows switch 1316 being connected to external device 1318. Data transmission between the switches and the external devices can be facilitated using any suitable data transmission media and/or standard. For example, data transmission can be based on fiber-optics or 10GBASE-T. As such, lower tier switches 1314 and 1316 can include circuitry and/or components (such as appropriate PRY chips, transceivers, edge ports, etc), to convert signals and enable data transmission with the external devices. Due to such configurations, switches 1314 and 1316 can each have a portion (e.g., half) of its ports configured based on the 10GBASE-KR standard and another portion (e.g., the other half) configured based on a different data transmission media and/or standard (e.g., fiber-optics, 10GBASE-T, etc.).
[0106] Through enabling data transmission between the middle and lower tier switches based on 10GBASE-KR, the middle and lower tier switches can be deployed closer to servers due to reduced size and thermal characteristics. In addition, because the standard, as discussed, is less costly to implement the cost of deploying the switches in the middle and lower tiers can be reduced. It should be noted that uplinks to the upper tier switches (which can be based on fiber-optics or other copper based transmission standards) can be located further away.
[0107] FIG. 14 illustrates an example configuration 1400 of a three-tier Clos-based network including a logical-tier 1/2/3 switch. In some embodiments, a software program executed by a processor of a network management server (not shown in FIG. 14) can operate the logical-tier 1/2/3 switch in a manner as to make the logical switch appear to other network devices as if it were a single unified switch. As shown in configuration 1400, there are effectively three layers of switches: an upper tier 1402, middle tier 1406, and lower tier 1410.
In some embodiments, data transmission over at least a portion of the network can be based on 10GBASE-KR operation. For example, various network components can include ASICs, copper traces, wires, and/or cables that are suitable for facilitating communication using the 10GBASE-KR standard. Those paths of the network communicating using 10GBASE-KR

are shown as dashed lines in FIG. 14.
[0108] The switches in the upper tier and the switches in the middle tier can be connected via a transpose box 1404. The manner in which the upper tier switches, middle tier switches, and transpose box 1404 are configured and operate can be similar to the manner described for the configuration and operation of the upper tier switches, middle tier switches, and transpose box illustrated in FIG. 12, except that the middle tier switches can be configured to operate based only on 10GBASE-KR (e.g., communication is not based on other transmission media and/or standards, such as fiber optics or 10GBASE-T). As further shown in configuration 1400, the switches in the middle tier and the lower tier can be connected via a set of transpose boxes 1408. The manner in which the switches in the middle tier and the switches in the lower tier are configured and operate can be similar to the manner described for the configuration and operation of middle tier switches 1308/1310, transpose box 1312 and lower tier switches 1314/1316 illustrated in FIG. 13, except that the middle tier switches, as discussed, can be configured to operate based only on 10GBASE-KR. In some embodiments, at least some of the switches in lower tier 1410 can include circuitry and/or components for directly connecting with servers, external switches, outside connections, etc.
(e.g., external switches 1412 and edge host 1414 shown in FIG. 14). For example, at least some of the switches in lower tier 1410 can include components and/or circuitry (such as appropriate PHY chips, transceivers, edge ports, etc) to convert signals and enable data transmission with external switches. Due to such configurations, at least some of the switches in lower tier 1410 can each have a portion (e.g., half) of its ports configured based on the 1.0GBASE-KR
standard and another portion (e.g., the other half) configured based on a different data transmission media and/or standard (e.g., fiber-optics, 10GBASE-T, etc.).
[0109] In some embodiments, example configuration 1400 can be of sufficient size as to be enclosed within a housing or chassis. More specifically, all cabling and/or wiring between the three tiers and the transpose boxes can be internal to a chassis and all data transmission can be efficiently run based on 10GBASE-KR. Through leveraging I OGBASE-KR and building a Clos network based on configuration 1400, a high-port count switch can be constructed using relatively cost effective and easily interchangeable parts.
101101 It should be understood that while the network and device configurations illustrated in FIGS. 11-14 show very specific implementations of enabling device connection and data transmission using, at least in part, 10GBASE-KR, many different configurations can be used to facilitate operation based on 10GBASE-KR. For example, the switches described in FIGS. 11-14 need not rely on edge connectors and transmission cables to connect to other devices. Any suitable wired and/or wireless connection mechanisms can be used, [0111] Furthermore, it should be understood that while many of the network, device, and connection configurations described herein facilitate data transmission based, in part, on 10GBASE-KR, any number of different transmission media and standards can be used. For example, the network and device configurations illustrated in FIGS. 11-14 can facilitate data transmission based on, in part, 10GBASE-KX4 or 400BASE-KR4 operation instead of 10GBASE-KR operation. As another example, the network and device configurations illustrated in FIGS. 11-14 can facilitate data transmission based on, in part, any future similar (e.g., short-distance) copper-based data transmission standard instead of operation.
[0112] Moreover, it should be understood that the various network configurations illustrated in FIGS. 11-14 can be deployed using any suitable process. For example, the network configurations can be deployed in a manner similar to the deployment process shown in FIG. 7.
[01131 Various embodiments can he described in view of the following clauses:
1. A data transmission network, comprising:
a first tier of network switches, each network switch capable of receiving and transmitting data over a network;
a second tier of network switches, each network switch in the second tier capable of receiving and transmitting data from at least one of the network switches in the first tier of network switches; and a network transpose box comprising a first group of network connectors and a second group of network connectors, at least a portion of the first group of network connectors each being connected to two or more of the second group of network connectors, at least a portion of the second group of network connectors each being connected to two or more of the first group of network connectors, wherein each of a subset of network switches of the first tier of network switches is connected to at least one of the first group of network connectors and each of a subset of network switches of the second tier of network switches is connected to at least one of the second group of network connectors, wherein a meshing of communication media connecting the first and second groups of network connectors implements a network topology wherein multiple switches of the first tier are each connected to multiple switches of second tier via the transpose box, and wherein at least a portion of the meshing of communication media facilitate the receiving and transmitting of data based on a backplane Ethernet standard.
2. The data transmission network of clause 1 wherein the backplane Ethernet standard is the 10GBASE-KR standard.
3. The data transmission network of clause 1 wherein at least one of the first tier of network switches is capable of receiving and transmitting data using a backplane Ethernet standard.
4. The data transmission network of clause 3 wherein the at least one of the first tier of network switches is also capable of receiving and transmitting data using a non-backplane Ethernet standard.
5. The data transmission network of clause 1 wherein the subset of network switches of the first tier of switches connected to the transpose box includes all of the switches of the first tier of switches, and wherein the subset of network switches of the second tier of switches connected to the transpose box includes all of the switches of the second tier of switches.
6. The data transmission network of clause 1 wherein at least one of the second tier of network switches is capable of receiving and transmitting data using a backplane Ethernet standard.
7. The data transmission network of clause 6 wherein the at least one of the second tier of network switches is also capable of receiving and transmitting data using a non-backplane Ethernet standard.
8. The data transmission network of clause 7 wherein the at least one of the second tier of network switches is capable of converting a data transmission between a backplane Ethernet standard and a non-backplane Ethernet standard.
9. The data transmission network of clause 1 further comprising a third tier of network switches, each network switch in the third tier capable of receiving and transmitting data from at least one of the second tier of network switches.
10. The data transmission network of clause 9 wherein the third tier of network switches is capable of receiving and transmitting data without using a backplane Ethernet standard.
11. The data transmission network of clause 1, wherein each of the subset of network switches of the first tier of network switches is connected to at least one of the first group of network connectors using a single transmission cable and each of the subset of network switches of the second tier of network switches is connected to at least one of the second group of network connectors using a single transmission cable.
12. The data transmission network of clause 9 further comprising a second transpose box comprising a third group of network connectors and a fourth group of network connectors, at least a portion of the third group of network connectors each being connected to two or more of the fourth group of network connectors, at least a portion of the fourth group of network connectors each being connected to two or more of the third group of network connectors, wherein each of a subset of network switches of the second tier of network switches is connected to at least one of the third group of network connectors and each of a subset of network switches of the third tier of network switches is connected to at least one of the fourth group of network connectors, wherein a meshing of communication media connecting the third and fourth groups of network connectors implements a network topology wherein multiple switches of the second tier are each connected to multiple switches of third tier via the transpose box, and wherein at least a portion of the meshing of communication media facilitate the receiving and transmitting of data based on a backplane Ethernet standard.
13. The data transmission network of clause 12 wherein each of the subset of network switches of the second tier of network switches is connected to at least one of the third group of network connectors using a single transmission cable and each of the subset of network switches of the third tier of network switches is connected to at least one of the fourth group of network connectors using a single transmission cable.
14. A data transmission system, the system comprising:
a network switch device configured to transmit and receive data according to a bad:plane Ethernet standard;
a network transpose box configured to transmit and receive data with the network switch device; and a transmission cable connecting the network switch device and the network transpose box, the transmission cable configured to facilitate the transmission of data between the network switch device and the network transpose box,
15. The system of clause 14 wherein the backplane Ethernet standard is the KR standard.
16. A method of deploying a data transmission network, comprising:
selecting a network topology for at least a portion of the data transmission network;
selecting one or more network devices to be deployed in the data transmission network, wherein the one or more network devices are configured to operate based on a backplanc Ethernet standard;
based at least in part upon the selected network topology, selecting a network transpose box including a plurality of network connectors connected using a plurality of communication media, the communication media being meshed in a way to implement the network topology, the network connectors being logically separated into a first group of network connectors and a second group of network connectors, at least a portion of the first group of network connectors each being connected to two or more of the second group of network connectors according to the selected topology;
connecting a first set of the one or more network devices to at least a portion of the network connectors of the first group and connecting a second set of the one or more network devices to at least a portion of the network connectors of the second group, each of the network devices being connected to at least one of the network connectors, wherein the first set of devices are able to communicate with the second set of devices according to the selected network topology.
17. The method of clause 16 wherein the backplane Ethernet standard is the KR standard.
18. The method of clause 16 wherein the backplane Ethernet standard is the KR4 standard.
19. The method of clause 16 wherein each of the network devices is connected to at least one of the network connectors using a single transmission cable.
20. A network switch device, comprising:
a physical support structure;
at least one network switching chip; and an edge connector supported by the physical support structure, the edge connector connected to the at least one network switching chip and configured to accept a transmission cable for transmitting a signal, wherein signal transmission is in accordance with a backplane Ethernet standard.
21. The network switch device of clause 20 wherein the backplane Ethernet standard is the 10GBASE-KR standard.
22. The network switch device of clause 20 wherein the network switch device is connected to a network transpose box via the transmission cable,
23. The network switch device of clause 20 wherein the edge connector is directly connected to a pin of the at least one network switching chip.
24. The network switch device of clause 20, wherein the network switch device further comprises circuitry, the circuitry being configured to transform the signal between an optical signal and an electrical signal.
25. A network transpose box, comprising:
a physical support structure;
a set of network connectors supported by the physical support structure, the set of network connectors separated into a first logical group and a second logical group each containing a plurality of the network connectors, each network connector configured to accept a transmission cable for transmitting a signal; and a plurality of connection media:
wherein the plurality of connection media facilitate signal transmission based on a backplane Ethernet standard, and wherein each instance of the connection media connecting one of the network connectors of the first logical group to one of the network connectors of the second logical group to enable a signal to be communicated between a transmission cable connected to the network connector of the first logical group and a transmission cable connected to the network connector of the second logical group, the connection media configured such that each network connector of the first group is connected to each network connector of the second group and each network connector of the second group is connected to each member of the first group.
26. The network transpose box of clause 25 wherein the backplane Ethernet standard is the 10GBASE-KR standard.
27. The network transpose box of clause 25 wherein the backplane Ethernet standard is the 40GBASE-KR4 standard.
28. The network transpose box of clause 25 wherein the plurality of connection media comprises a plurality of printed circuit board traces.
29. The network transpose box of clause 25 wherein two or more network connectors are each connected to a network device, and wherein each network device connected to a network connector of the first group is capable of being connected to each network device connected to at least two of the network connectors of the second group.
30. The network transpose box of clause 29 wherein each of the set of network switch devices is connected to the transpose box using a number of transmission cables that is less than the number of ports to be connected to the transpose box.
[0114] As discussed above, the various embodiments can be implemented in a wide variety of operating environments, which in some cases can include one or more user computers, computing devices, or processing devices which can be used to operate any of a number of applications. User or client devices can include any of a number of general purpose personal computers, such as desktop or laptop computers running a standard operating system, as well as cellular, wireless, and handheld devices running mobile software and capable of supporting a number of networking and messaging protocols. Such a system also can include a number of workstations running any of a variety of commercially-available operating systems and other known applications for purposes such as development and database management. These devices also can include other electronic devices, such as dummy terminals, thin-clients, gaming systems, and other devices capable of communicating via a network.
[0115] Various aspects also can be implemented as part of at least one service or Web service, such as may be part of a service-oriented architecture. Services such as Web services can communicate using any appropriate type of messaging, such as by using messages in extensible markup language (XML) format and exchanged using an appropriate protocol such as SOAP (derived from the "Simple Object Access Protocol"). Processes provided or executed by such services can be written in any appropriate language, such as the Web Services Description Language (WSDL). Using a language such as WSDL allows for functionality such as the automated generation of client-side code in various SOAP
frameworks.
[0116] Most embodiments utilize at least one network that would be familiar to those skilled in the art for supporting communications using any of a variety of commercially-available protocols, such as TCP/IP, OSI, FTP, UPnP, NFS, and CIES. The network can be, for example, a local area network, a wide-area network, a virtual private network, the Internet, an intranet, an extranet, a public switched telephone network, an infrared network, a wireless network, and any combination thereof.
[01171 In embodiments utilizing a Web server, the Web server can run any of a variety of server or mid-tier applications, including HTTP servers, FTP servers, CG1 servers, data servers, Java servers, and business application servers. The server(s) also may be capable of executing programs or scripts in response requests from user devices, such as by executing one or more Web applications that may be implemented as one or more scripts or programs written in any programming language, such as Java, C, C# or C++, or any scripting language, such as Perl, Python, or TCL, as well as combinations thereof. The server(s) may also include database servers, including without limitation those commercially available from Oracle", Microsoft , Sybasee, and IBM
[0118] The environment can include a variety of data stores and other memory and storage media as discussed above. These can reside in a variety of locations, such as on a storage medium local to (and/or resident in) one or more of the computers or remote from any or all of the computers across the network. In a particular set of embodiments, the information may reside in a storage-area network ("SAN") familiar to those skilled in the art.
Similarly, any necessary files for performing the functions attributed to the computers, servers, or other network devices may be stored locally and/or remotely, as appropriate. Where a system includes computerized devices, each such device can include hardware elements that may be electrically coupled via a bus, the elements including, for example, at least one central processing unit (CPU), at least one input device (e.g., a mouse, keyboard, controller, touch screen, or keypad), and at least one output device (e.g., a display device, printer, or speaker).
Such a system may also include one or more storage devices, such as disk drives, optical storage devices, and solid-state storage devices such as random access memory ("RAM") or read-only memory ("ROM"), as well as removable media devices, memory cards, flash cards, etc.
[0119] Such devices also can include a computer-readable storage media reader, a communications device (e.g., a modem, a network card (wireless or wired), an infrared communication device, etc.), and working memory as described above. The computer-readable storage media reader can be connected with, or configured to receive, a computer-readable storage medium, representing remote, local, fixed, and/or removable storage devices as well as storage media for temporarily and/or more permanently containing, storing, transmitting, and retrieving computer-readable information. The system and various devices also typically will include a number of software applications, modules, services, or other elements located within at least one working memory device, including an operating system and application programs, such as a client application or Web browser. It should be appreciated that alternate embodiments may have numerous variations from that described above. For example, customized hardware might also be used and/or particular elements might be implemented in hardware, software (including portable software, such as applets), or both. Further, connection to other computing devices such as network input/output devices may be employed.
[0120] Storage media and computer readable media for containing code, or portions of code, can include any appropriate media known or used in the art, including storage media and communication media, such as but not limited to volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage and/or transmission of information such as computer readable instructions, data structures, program modules, or other data, including RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, solid state drives (SSD) which use solid state flash memory like Single-Level Cell (SLC) and Multi-Level Cell (MLC),or any other medium which can be used to store the desired information and which can be accessed by the a system device. Based on the disclosure and teachings provided herein, a person of ordinary skill in the art will appreciate other ways and/or methods to implement the various embodiments.
[0121] The specification and drawings are, accordingly, to be regarded in an illustrative rather than a restrictive sense. It will, however, be evident that various modifications and changes may be made thereunto without departing from the broader spirit and scope of the invention as set forth in the claims.

Claims (28)

  1. CLAIMS:
    A data transmission network, comprising:
    a first tier of network spine switches, each network spine switch in the first tier configured to receive and transmit data over a network;
    a second tier of network edge switches, each network edge switch in the second tier configured to receive and transmit data from at least one of the network spine switches in the first tier of network spine switches; and a network transpose box comprising a first group of network connectors associated with the first tier of network spine switches, a second group of network connectors associated with the second tier of network edge switches, and a meshing of communication media, the meshing of communication media being configured to implement a network topology that physically couples each one of the network connectors of the first group of network connectors individually to each one of the network connectors of the second group of network connectors to enable data transmission from the first tier of network spine switches to the second tier of network edge switches based on a backplane Ethernet standard, wherein a first network spine switch of the first tier of network spine switches is connected to a first network connector of the first group of network connectors of the transpose box via a single first cable, the first network spine switch being configured to transmit data with each one of the network edge switches in the second tier of network edge switches through the meshing of communication media.
  2. 2. The data transmission network of claim 1 wherein the backplane Ethernet standard is a 10GBASE-KR standard.
  3. 3. The data transmission network of claim I wherein at least one of the first tier of network spine switches is configured to receive and transmit data using a backplane Ethernet standard
  4. 4. The data transmission network of claim 3 wherein the at least one of the first tier of network spine switches is also configured to receive and transmit data using a non-backplane Ethernet standard.
  5. 5. The data transmission network of claim 1 wherein at least one of the second tier of network edge switches is configured to receive and transmit data using a backplane Ethernet standard
  6. 6. The data transmission network of claim 5 wherein the at least one of the second tier of network edge switches is also configured to receive and transmit data using a non-backplane Ethernet standard.
  7. 7. The data transmission network of claim 6 wherein the at least one of the second tier of network edge switches is configured to convert a data transmission between a backplane Ethernet standard and a non-backplane Ethernet standard.
  8. 8. The data transmission network of claim 1 further comprising a third tier of network switches, each network switch in the third tier configured to receive and transmit data from at least one of the second tier of network edge switches.
  9. 9 The data transmission network of claim 8 wherein the third tier of network switches is configured to receive and transmit data without using a backplane Ethernet standard.
  10. The data transmission network of claim 8 further comprising a second network transpose box comprising a third group of network connectors and a fourth group of network connectors, at least a portion of the third group of network connectors each being connected to two or more of the fourth group of network connectors, at least a portion of the fourth group of network connectors each being connected to two or more of the third group of network connectors, wherein each of a second subset of network switches of the second tier of network edge switches is connected to at least one of the third group of network connectors and each of a subset of network switches of the third tier of network switches is connected to at least one of the fourth group of network connectors, wherein a meshing of communication media that is located within the network transpose box connects the third and fourth groups of network connectors and implements a network topology wherein multiple network switches of the second tier are each connected to multiple network switches of the third tier via the second network transpose box, and wherein at least a portion of the meshing of communication media facilitate receiving and transmitting of data based on a backplane Ethernet standard.
  11. 11. The data transmission network of claim 10 wherein each of the second subset of network switches of the second tier of network switches is connected to at least one of the third group of network connectors using a respective transmission cable and each of the subset of network switches of the third tier of network switches is connected to at least one of the fourth group of network connectors using a single transmission cable.
  12. 12. A data transmission system, the system comprising:
    a first tier of network spine switch devices;
    a second tier of network edge switch devices, the first tier of network spine switch devices and the second tier of network edge switch devices configured to transmit and receive data according to a backplane Ethernet standard; and a network transpose box configured to transmit data from the first tier of network spine switch devices to the second tier of network edge switch devices through a meshing of communication media, the network transpose box comprising:
    a first group of network connectors associated with the first tier of network spine switch devices, a second group of network connectors associated with the second tier of network edge switch devices, and a meshing of communication media, the meshing of communication media appearing as a single unified switch to one or more other network devices, wherein the meshing of communication media physically connects each of the first group of network connectors individually to each of the second group of network connectors and implements a network topology that connects, using a single transmission cable for each connection, multiple network spine switch devices of the first tier of network spine switch devices to multiple network edge switch devices of the second tier of network edge switch devices via the single transmission cable between each spine switch device of the first tier and each edge switch device of the second tier inside of the network transpose box, wherein a first network spine switch device of the first tier of network spine switch devices is connected to a first network connector of the first group of network connectors of the transpose box via a single first cable, the first network spine switch device being configured to transmit data with each one of the network edge switch devices in the second tier of network edge switch devices through the meshing of communication media.
  13. 13 The data transmission system of claim 12 wherein the backplane Ethernet standard is a 10GBASE-KR standard.
  14. 14. A method of deploying a data transmission network, comprising:
    selecting a network topology for at least a portion of the data transmission network;
    selecting one or more network devices to be deployed in the data transmission network, wherein the one or more network devices are configured to operate based at least in part on a backplane Ethernet standard;
    based at least in part upon the selected network topology, selecting a network transpose box, the network transpose box including network connectors separated into a first group of network connectors associated with a first tier of network spine devices and a second group of network connectors associated with a second tier of network edge devices, and a plurality of communication media that is located within the network transpose box, the plurality of communication media having a meshing of communication media that implements the network topology that physically connects, using a single cable for each connection, multiple network spine devices of the first tier of network spine devices individually to multiple network edge devices of the second tier of network edge devices via the single cable between each spine device of the first tier and each edge device of the second tier inside of the network transpose box, the first group of network connectors each being connected to each of the second group of network connectors according to the selected topology using the meshing of' communication media, the meshing of communication media appearing as a single unified switch to one or more other network devices; and connecting a first network spine device of the first tier of network spine devices to a first network connector of the first group of network connectors of the transpose box via a single first cable, the first network device being configured to communicate with each one of the network edge devices in the second tier of network edge devices through the meshing of communication media according to the selected network topology.
  15. 15 The method of claim 14 wherein the backplane Ethernet standard is a 10GBASE-KR standard.
  16. 16. The method of claim 14 wherein the backplane Ethernet standard is a 40GBASE-KR4 standard.
  17. 17. The method of claim 14 wherein each of the network devices is connected to at least one of the network connectors using a respective transmission cable.
  18. 18. The data transmission system of claim 12, the network switch devices further comprising:
    a physical support structure;
    at least one network switching chip; and an edge connector supported by the physical support structure, the edge connector connected to the at least one network switching chip and configured to accept a transmission cable for transmitting a signal, wherein signal transmission is in accordance with a backplane Ethernet standard
  19. 19. The data transmission system of claim 18 wherein the backplane Ethernet standard is a 10GBASE-KR standard.
  20. 20. The data transmission system of claim 18 wherein the network switch devices are connected to the network transpose box via the transmission cable.
  21. 21. The data transmission system of claim 18 wherein the edge connector is directly connected to a pin of the at least one network switching chip.
  22. 22. The data transmission system of claim 18, wherein the network switch devices further comprise circuitry, the circuitry being configured to transform the signal between an optical signal and an electrical signal.
  23. 23. A network transpose box comprising:
    a physical support structure;
    a set of network connectors supported by the physical support structure, the set of network connectors separated into a first logical group and a second logical group each containing a plurality of the network connectors, each network connector configured to accept a transmission cable for transmitting a signal; and a plurality of connection media that is located within the network transpose box, wherein the plurality of connection media facilitate signal transmission based on a backplane Ethernet standard, wherein each instance of the connection media connecting each of the network connectors of the first logical group to each of the network connectors of the second logical group enables a signal to be communicated between a transmission cable connected to any network connector of the first logical group and a transmission cable connected to any network connector of the second logical group, the connection media configured such that each network connector of the first logical group is physically connected individually to each network connector of the second logical group and each network connector of the second logical group is physically connected individually to each member of the first logical group, wherein at least some of a first tier of network spine switches is connected to at least one associated network connector of the first logical group of network connectors within the network transpose box using a single transmission cable and at least some of a second tier of network edge switches is connected to at least one associated network connector of the second logical group of network connectors using a single transmission cable within the network transpose box, and wherein the network transpose box is configured to transmit data from the first tier of network edge switches to the second tier of network edge switches through the plurality of connection media that is located within the network transpose box, and wherein the first logical group and second logical group appears to other network devices as a single unified logical group.
  24. 24 The network transpose box of claim 23 wherein the backplane Ethernet standard is a 10GBASE-KR standard.
  25. 25. The network transpose box of claim 23 wherein the backplane Ethernet standard is a 40GBASE-KR4 standard.
  26. 76. The network transpose box of claim 23 wherein the plurality of connection media comprises a plurality of printed circuit board traces.
  27. 27. The network transpose box of claim 23 wherein two or more network connectors are each connected to a network device, and wherein each network device connected to a network connector of the first logical group is configured to connect to each network device connected to at least two of the network connectors of the second logical group.
  28. 28. The network transpose box of claim 27 wherein each of the set of network switch devices is connected to the transpose box using a number of transmission cables that is less than the number of ports to be connected to the network transpose box.
CA2831607A 2011-03-29 2012-03-28 Network transpose box and switch operation based on backplane ethernet Active CA2831607C (en)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US13/074,998 2011-03-29
US13/074,998 US20120250679A1 (en) 2011-03-29 2011-03-29 Network Transpose Box and Switch Operation Based on Backplane Ethernet
PCT/US2012/030972 WO2012135361A1 (en) 2011-03-29 2012-03-28 Network transpose box and switch operation based on backplane ethernet

Publications (2)

Publication Number Publication Date
CA2831607A1 CA2831607A1 (en) 2012-10-04
CA2831607C true CA2831607C (en) 2018-07-17

Family

ID=46927204

Family Applications (1)

Application Number Title Priority Date Filing Date
CA2831607A Active CA2831607C (en) 2011-03-29 2012-03-28 Network transpose box and switch operation based on backplane ethernet

Country Status (7)

Country Link
US (1) US20120250679A1 (en)
EP (1) EP2691866B1 (en)
JP (1) JP5792372B2 (en)
CN (1) CN103608797B (en)
CA (1) CA2831607C (en)
SG (1) SG194018A1 (en)
WO (1) WO2012135361A1 (en)

Families Citing this family (40)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8621111B2 (en) 2010-09-22 2013-12-31 Amazon Technologies, Inc. Transpose box based network scaling
US9363208B1 (en) 2011-03-29 2016-06-07 Amazon Technologies, Inc. Logical switches
US9008510B1 (en) 2011-05-12 2015-04-14 Google Inc. Implementation of a large-scale multi-stage non-blocking optical circuit switch
US8862537B1 (en) * 2011-06-30 2014-10-14 Sumo Logic Selective structure preserving obfuscation
US9066160B2 (en) * 2011-07-07 2015-06-23 Alcatel Lucent Apparatus and method for protection in a data center
US8902751B1 (en) * 2012-01-26 2014-12-02 Google Inc. Multi-stage switching topology
US20130250802A1 (en) * 2012-03-26 2013-09-26 Praveen Yalagandula Reducing cabling costs in a datacenter network
US9020346B2 (en) * 2012-09-11 2015-04-28 Inphi Corporation Optical communication interface utilizing coded pulse amplitude modulation
US9197324B1 (en) 2012-04-09 2015-11-24 Inphi Corporation Method and system for transmitter optimization of an optical PAM serdes based on receiver feedback
US9473422B1 (en) * 2012-05-09 2016-10-18 Google Inc. Multi-stage switching topology
US8983291B1 (en) 2012-07-30 2015-03-17 Inphi Corporation Optical PAM modulation with dual drive mach zehnder modulators and low complexity electrical signaling
US9647799B2 (en) 2012-10-16 2017-05-09 Inphi Corporation FEC coding identification
US9787567B1 (en) 2013-01-30 2017-10-10 Big Switch Networks, Inc. Systems and methods for network traffic monitoring
US9432123B2 (en) 2013-03-08 2016-08-30 Inphi Corporation Adaptive mach zehnder modulator linearization
US9240957B2 (en) * 2013-04-16 2016-01-19 Dell Products L.P. System and method for symmetric switching fabrics
US9258253B2 (en) 2013-05-14 2016-02-09 Dell Products L.P. System and method for flexible switching fabric
US9960878B2 (en) * 2013-10-01 2018-05-01 Indian Institute Of Technology Bombay Scalable ultra dense hypergraph network for data centers
US10498570B2 (en) 2013-10-02 2019-12-03 Inphi Corporation Data communication systems with forward error correction
US9183104B2 (en) * 2013-12-20 2015-11-10 International Business Machines Corporation Validating connection, structural characteristics and positioning of cable connectors
US9654852B2 (en) * 2013-12-24 2017-05-16 Nec Corporation Scalable hybrid packet/circuit switching network architecture
WO2015152940A1 (en) 2014-04-04 2015-10-08 Hewlett-Packard Development Company, L.P. Flexible input/output zone in a server chassis
US9634959B2 (en) 2014-04-08 2017-04-25 International Business Machines Corporation High-density, fail-in-place switches for computer and data networks
US10270645B2 (en) * 2014-07-21 2019-04-23 Big Switch Networks, Inc. Systems and methods for handling link aggregation failover with a controller
US9813323B2 (en) 2015-02-10 2017-11-07 Big Switch Networks, Inc. Systems and methods for controlling switches to capture and monitor network traffic
US11178072B2 (en) * 2015-06-10 2021-11-16 Mobileye Vision Technologies Ltd. Image processor and methods for processing an image
US11277217B2 (en) 2015-06-30 2022-03-15 Ciena Corporation Flexible Ethernet switching systems and methods
US9800361B2 (en) * 2015-06-30 2017-10-24 Ciena Corporation Flexible ethernet switching systems and methods
WO2017101114A1 (en) 2015-12-18 2017-06-22 华为技术有限公司 Networking method for datacenter network and datacenter network
US10425358B2 (en) * 2016-09-29 2019-09-24 International Business Machines Corporation Network switch architecture supporting multiple simultaneous collective operations
US10158929B1 (en) * 2017-02-17 2018-12-18 Capital Com SV Investments Limited Specialized optical switches utilized to reduce latency in switching between hardware devices in computer systems and methods of use thereof
CN108879150B (en) * 2017-05-15 2020-12-15 华为技术有限公司 Cable backboard and communication equipment
US10419327B2 (en) 2017-10-12 2019-09-17 Big Switch Networks, Inc. Systems and methods for controlling switches to record network packets using a traffic monitoring network
CN111277519A (en) * 2020-02-21 2020-06-12 苏州浪潮智能科技有限公司 Exchange chip extension system and switch
CN111786913A (en) * 2020-09-04 2020-10-16 苏州浪潮智能科技有限公司 Server cabinet and wireless node interconnection system architecture thereof
TW202241093A (en) * 2021-04-01 2022-10-16 智邦科技股份有限公司 Scaling network fabric size system
US11870682B2 (en) 2021-06-22 2024-01-09 Mellanox Technologies, Ltd. Deadlock-free local rerouting for handling multiple local link failures in hierarchical network topologies
US11765103B2 (en) * 2021-12-01 2023-09-19 Mellanox Technologies, Ltd. Large-scale network with high port utilization
US20240171886A1 (en) 2022-11-17 2024-05-23 Panduit Corp. Fabric modules for high-radix networks
US20240171887A1 (en) 2022-11-17 2024-05-23 Panduit Corp. Fabric modules for server to switch connections
US20240168234A1 (en) 2022-11-17 2024-05-23 Panduit Corp. Fabric network modules

Family Cites Families (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0549052A (en) * 1991-08-14 1993-02-26 Takenaka Komuten Co Ltd In-premise communication system and repeater used for the system
JPH07288853A (en) * 1994-04-15 1995-10-31 Mitsubishi Electric Corp Wiring system
GB0320168D0 (en) * 2003-08-28 2003-10-01 Univ Manchester A sensing device
US7298946B2 (en) * 2004-12-22 2007-11-20 Hewlett-Packard Development Company, L.P. Multi-fiber cable for efficient manageability of optical system
US7389046B1 (en) * 2005-01-18 2008-06-17 Woven Systems, Inc. Simplified switching interconnect using wavelength division multiplexing
US7340363B2 (en) * 2005-02-01 2008-03-04 Alpha Networks Inc. Apparatus of automatically detecting type of cable connected to network and method thereof
ATE544266T1 (en) * 2005-06-10 2012-02-15 Partec Cluster Competence Ct Gmbh DATA COMMUNICATION METHOD AND SYSTEM WITH VIRTUAL PORTS
US20100025470A1 (en) * 2006-05-03 2010-02-04 Katsuya Nakagawa Bar code reader, optical mouse, and bar code reading method
US20100115306A1 (en) * 2008-11-05 2010-05-06 Wael William Diab Method and system for control of energy efficiency and associated policies in a physical layer device
US8327487B2 (en) * 2008-01-31 2012-12-11 Black & Decker Inc. Vacuum filter cleaning device
US8761152B2 (en) * 2008-10-14 2014-06-24 William Marsh Rice University Method and system for scalable ethernet
US8370704B2 (en) * 2009-03-09 2013-02-05 Intel Corporation Cable interconnection techniques
US8270830B2 (en) * 2009-04-01 2012-09-18 Fusion-Io, Inc. Optical network for cluster computing
TWI430735B (en) * 2009-07-21 2014-03-11 Qisda Corp Cover and electronic apparatus
US9256560B2 (en) * 2009-07-29 2016-02-09 Solarflare Communications, Inc. Controller integration
US8661309B2 (en) * 2010-01-29 2014-02-25 Broadcom Corporation Systems for high-speed backplane applications using pre-coding
US8621111B2 (en) * 2010-09-22 2013-12-31 Amazon Technologies, Inc. Transpose box based network scaling

Also Published As

Publication number Publication date
CN103608797B (en) 2019-07-05
JP2014509819A (en) 2014-04-21
SG194018A1 (en) 2013-11-29
JP5792372B2 (en) 2015-10-14
CN103608797A (en) 2014-02-26
EP2691866A1 (en) 2014-02-05
US20120250679A1 (en) 2012-10-04
CA2831607A1 (en) 2012-10-04
WO2012135361A1 (en) 2012-10-04
EP2691866B1 (en) 2019-07-31
EP2691866A4 (en) 2014-12-24

Similar Documents

Publication Publication Date Title
CA2831607C (en) Network transpose box and switch operation based on backplane ethernet
US9813355B2 (en) Logical switches
US9330052B2 (en) Transpose box based network scaling
CA2811236C (en) Transpose boxes for network interconnection
US9106564B2 (en) Incremental high radix network scaling
US8934483B2 (en) Data center switch
US9750135B2 (en) Dual faced ATCA backplane
Insensitive MULTI-INTERFACE POINT-TO-POINT SWITCHING SYSTEM (MIPPSS) 5 HAVING AN INTERNAL UNIVERSAL SIGNAL FORMAT

Legal Events

Date Code Title Description
EEER Examination request

Effective date: 20130926